What: BCG's Applied AI Index 2026, a survey of 1,330 executives, finds almost half of companies now generate measurable value from AI, with 7.5% of them (the "future-built") capturing most of it.

Why it is important: The report ties AI value to organisational change, not tools, with nearly three-quarters of future-built companies' top priorities sitting in people, organisation and processes. That matches the retail evidence in the collection that execution, not technology, is the barrier.

BCG's Applied AI Index 2026 surveyed 1,330 executives across more than 20 sectors. Almost half of companies now generate measurable value from AI, but most of it sits with a small group of "future-built" companies, which are 7.5% of the sample, up from 5% last year. They generate 2.4 times as much top-line growth as laggards and three times the EBITDA growth rate.
BCG attributes the gap to strategic clarity combined with applied AI. Future-built companies run a multiyear AI program, and 95% track value on the P&L or a proxy. Across the sample, AI spending has reached 3.3% of revenue, mostly outside IT budgets. Agentic AI's share of AI value rose from 17% to 22% and is projected to reach 39% by 2030, yet only 5% of companies have all six agent controls in place.
Future-built companies put nearly three-quarters of their top priorities into people, organisation and processes, ahead of technology and algorithms. Companies expect AI to cut the workforce by roughly 10% to 15%, concentrated in middle management and frontline leadership. Retail scores 49 out of 100 on AI maturity, and 48% of retailers are future-built or scaling.

IADS Notes: The BCG report's findings land on a retail sector that is adopting AI unevenly and struggling to turn it into returns. In the UK, wholesale and retail sits near the bottom of sector adoption rankings, with 15.7% of businesses using at least one AI technology (Retail Week, July 2026). In the U.S., AI investment among retail and fashion companies is rising, but adoption remains broad and shallow, and value depends on redesigning workflows, governance and data (WWD, July 2026). BCG's own CEO survey shows the same execution gap: nearly nine in ten CEOs report AI benefits in targeted areas, but only 14% have defined the financial impact of every AI project (BCG, July 2026). The agent control gap is already visible in retail operations, where agents able to issue refunds, adjust loyalty points or create purchase orders need a named business owner, scoped permissions and audit logs (RH-ISAC, August 2026). On the workforce side, AI initiatives most often stall in how middle managers translate mandates into daily practice (Harvard Business Review, September 2026).

THE APPLIED AI INDEX 2026 : The Formula for Agentic AI Value

PRINTABLE VERSION HERE  

Beauty in department stores offers quite a uniform picture: the first thing shoppers smell and see, sitting on the ground floor to drive traffic and conversion. Despite this long-standing model, the category is challenged as department stores are no longer the primary channel for beauty discovery. Competition from speciality retailers, e-commerce, and social media has eroded their historical leadership. For example, only 24% of US shoppers now visit a department store to learn about new products, with 71% turning to Sephora or Ulta Beauty, showing that consumers want more than traditional brand-counter organisations. The proliferation of discounts and new brands also brings additional challenges.

However, despite that bleak situation, many department stores recently invested in their beauty floors, from John Lewis to Macy’s, to benefit from the current category momentum. Indeed, in 2023, global beauty market retail sales grew to $446 billion, up 10% from 2022. In 2025, McKinsey expected beauty to continue growing at 5% over the next five years. The outlook is still positive in 2026, and the category should reach $590 billion by 2030. In that picture, department stores have the weakest outlook, having already declined by approximately 2% annually since 2019 and facing a similar outlook in the coming years. So, is beauty still worth it for department stores, or could it go extinct? If so, what could replace it and drive similar traffic and conversion?

The new beauty floor: assortment, services and productivity 

Product categories to drive growth 

For now, beauty remains a key category, though category leadership has shifted in recent years. According to McKinsey, fragrance has overtaken colour cosmetics as the third-largest beauty category. It is expected to continue growing across price points, driven by new behaviours such as using different fragrances for different occasions. This growth is supported by masstige brands as well as prestige and niche scents, as exemplified by the recent addition of new brands such as Argentinian brand Fueguia at Galeries Lafayette Haussmann, or the renovation of Selfridges’ fragrance hall. Overall, as stated by most IADS members, there is no sign of slowing down in niche fragrances.

On the other hand, skincare is at a crossroads. With inflation and information at hand, consumers are less convinced that higher prices automatically mean better performance, as dermatologist-backed, K-beauty and dermo-cosmetics brands show customers that lower-priced offerings can achieve similar results. In parallel, as IADS members note, the traditional anti-ageing skincare promise is being challenged by the notion of longevity, a more holistic perspective that includes supplements, wellness, and aesthetic treatments.

The colour cosmetics category has been under pressure after several years of minimalist makeup trends, but now shows renewed interest in bolder makeup looks and hybrid makeup-skincare products.

In addition, IADS members mentioned that beauty is expanding into adjacent categories: hair care and specifically scalp care, supplements, sleep care, nutrition, sun care, nail care, aesthetic services, and at-home devices such as LED masks and radio-frequency tools. Finally, men’s beauty is growing across markets according to IADS members, with brands such as Nivea For Men, Biotherm Homme and Horace.

Making beauty services visible and valuable 

Access to the product categories above is a given for customers. What makes a difference is the customer experience, including services. Mostly in treatment cabins, services are hidden for obvious reasons of intimacy. Usually managed by brands, they are sometimes not even advertised in the store, making them invisible to customers. Department store executives and brands should challenge their location and visibility, as services could serve as a floor-animation tool, as shown by department stores’ wellness initiatives.

First on the trend, Harrods opened their Wellness Clinic in 2017, an appointment-only 1,000 sqm space on the fourth floor in Knightsbridge. Closed-door, yet advertised, the space was never intended to be a traditional spa or beauty salon. Harrods envisioned it as a high-end, results-driven clinic offering a curated selection of treatments designed not for pampering, but rather for measurable results. The range is unusually wide for a department store, covering everything from injectables and body contouring to cryotherapy, DNA-based personalised skincare, IV vitamin infusions, dental services, osteopathy, podiatry, physiotherapy, personal training, nutritional coaching, and more, making personalised service a reason to visit. Second, with a different approach, Galeries Lafayette opened their Wellness Galerie in 2022, with a full 3,000 sqm floor, including an unprecedented 60% space share given to highly visible services: a gym, infrared sauna, and a hyperbaric chamber, to name a few. Some services are now discontinued, though, showing how challenging it is to make those offerings successful and profitable compared to product-selling core activities.

Fewer SKUs, more cash 

In a world where discount is king, it is no longer worth carrying some brands, as department stores can never compete on prices. This is the case for premium fragrances for some IADS members, for example. Instead, shifting to niche labels can act as a footfall and growth driver. This comes down to the department store role in discovery rather than offering the broadest assortment.

For some, the game is on cash. Challenging brands to reduce quantities in each SKU (Estée Lauder’s Double Wear 54 shades of foundation are certainly not selling equally), down to individual lipstick or foundation lines, has unlocked cash savings without degrading how customers perceive assortment breadth. In that case, stores are graded A, B, C, D with quantities aligned to the grade. This is typically a beauty‑specific problem, underscoring how being an efficient merchant includes negotiating with brands to reduce the number of SKUs carried. This is no easy task, though, as department stores must ensure customers still believe they are offered a broad, diversified assortment. Department stores must manage this contradiction: “simulate” a broad yet curated brand and product assortment while rationalising SKUs and quantities.

Bringing beauty to another level, or dimension: space, placement and the question of digital

The need for renewed differentiation 

With inflation and the proliferation of information, consumers are more value-conscious and knowledgeable. As a result, they are increasingly buying across price points, mixing mass and prestige products within the same routine. For example, around 20% of brands carried by US speciality beauty retailers are also sold at Walmart, illustrating how price and channel boundaries are blurring. As assortments become more similar across retailers, exclusive and differentiated products become competitive advantages. Without a differentiated assortment or shopping experience, consumers choose the retailer offering price, convenience, delivery speed or loyalty benefits, especially for replenishment purchases where loyalty to a retailer doesn’t matter. To differentiate and make a trip to the store worth the effort, department stores are creating alternative offerings through multi-brand areas selling small, niche brands or French pharmacy spaces, as exemplified by Falabella and Galeries Lafayette. 

Borrowing from Sephora: the agnostic model 

Sephora and Ulta Beauty had trained consumers to expect an agnostic experience with open shelves and independent expert advice. Even though department store beauty counters as we know them cannot replicate this agnostic model, building new experiential spaces has become a growing concern.

Falabella gives the most advanced answer to this problem with Beauty F, its own specialist beauty format, designed, built and operated fully in-house: an enclosed space with a single entry point within Falabella department stores that functions as a standalone retail business, not a collection of brand concessions. The open-shelf layout integrates testing stations and advisory zones. The sales force consists entirely of Falabella employees, avoiding competing loyalties. Falabella deals directly with brands, importing and curating the selection, including younger, trend-driven, celebrity brands and K-beauty brands, and supports CapEx costs.

One of the most commercially significant Beauty F successes is re-engaging customers who haven’t shopped in more than 12 months. Twenty per cent of the customers are aged 18-35, a cohort that has largely abandoned department stores. According to Falabella, the share of young customers and new customer acquisition have both doubled since the format launched, and new brands introduced through the format contribute significantly more than the legacy assortment.

Beauty as a vertical journey 

Another differentiating option exemplified by Galeries Lafayette is to spread beauty across three floors, a deliberate vertical journey. The -1 floor offers an updated, more realistic version of the Wellness Galerie initially launched in 2022, oriented towards care brands and beauty services. The ground floor centres on fragrance and makeup, including recent additions such as Victoria Beckham, Louis Vuitton makeup and 13 niche fragrance brands. Creating a destination within the luxury RTW floor, the +1 level brings together luxury fashion, jewellery and beauty, with high-end perfumery spaces for Loewe, Dior, Chanel, Maison Francis Kurkdjian, Officine Universelle Buly 1803 and more.

This overhaul aims to differentiate Galeries Lafayette from competitors by blending luxury and beauty and giving the luxury RTW floor a more lifestyle feel. Now accounting for about 10% of the store’s annual volume, beauty has posted double-digit growth in 2025, continuing into 2026. The beauty selling space has expanded to 3,700 sqm, an increase of 500 sqm, bringing the total brand count to 440. This demonstrates that, despite a bleak outlook, department stores can still play their cards right.

Beauty growth is moving online, TikTok accelerating the shift 

E-commerce is expected to drive most of the beauty sales growth through 2030. Globally, e-commerce already represents the biggest channel at 28% of beauty sales, ahead of grocery at 19%. The US market illustrates this. Amazon, Ulta Beauty, Sephora and TikTok together account for just under half of US beauty sales. Amazon and TikTok’s combined beauty sales could soon exceed Sephora and Ulta combined sales. TikTok is a complete commercial channel, as US beauty sales on the platform have grown at approximately 260% annually since 2023. This is not a niche youth phenomenon, as the 35-55+ bracket makes up 49% of daily users.

Department stores are not built for the content TikTok demands. The format that wins on TikTok Shop is high-frequency and creator-native, a volume and style of content most legacy retailers are not resourced to produce, and simply repurposing existing brand assets does not work. Department stores don’t have the beauty-native content muscle of Sephora (nor the lower cost structure that lets them channel more funds to online), and maintaining a content presence for every brand on the floor is not viable at department store scale. A cheaper way to close part of this gap is to translate viral discovery into physical offerings. This is what Sephora has built with their ‘hot on social media’ in-store visibly labelled spaces, a feature department stores could take cues from. With its constantly changing beauty ‘newsroom’ in-store sections, Mecca is also a beauty retailer to watch.

In that digital-first environment, department stores have steadily ceded beauty share to online retailers and speciality players with strong e-commerce operations. To defend their positions, IADS members have strengthened their online operations, resulting in modest growth in average share compared to competitors, from 14% of the beauty business in 2024 to 16% in 2025. Some, such as El Palacio de Hierro, set up dedicated beauty accounts on Instagram and TikTok, knowing social media economics remain challenging because of promotions, affiliate commissions, and high customer acquisition costs.

If not beauty, then what?

The easy option: more products 

What could replace beauty with equivalent traffic potential? The temptation to fill the ground-floor, highly visible space with more products would be strong. Creating or expanding the number and surface area of pop-up spaces to feature commercial exhibitions comparable to what Le Bon Marché offers, or these spaces to brands for rotating product discovery, is an option. Another option is to offer larger concession spaces, especially to luxury brands, allowing them to present their full brand vision in one space, such as a free-standing store. Though both solutions would generate short-term revenue, they would transform the ground floor into a mall-like environment and compromise the store's identity. Besides, none of these options would avoid e-commerce cannibalisation unless spaces are filled only with exclusive products. Non-product-selling businesses would avoid that, though they raise the question of how to achieve the same commercial results as product selling. 

F&B as the new footfall driver 

F&B offerings present opportunities, as food experiences drive significant dwell time, repeat traffic and purchases for food replenishment. Harrods measured that when customers engage with the store’s restaurants and bars, they spend twice as long in the building and twice as much money. Department stores already have F&B options, but the goal here would be to make it a true destination for meals and food purchases, such as the La Grande Épicerie de Paris store in Paris, or the food floors at KaDeWe in Berlin and at Isetan in Tokyo Shinjuku. But it raises the question of how to push customers to other floors for cross-category shopping. So, replacing beauty with F&B would probably require more than repurposing a floor. It would require an entire store reshuffle to include more food offerings across floors.

Selling entertainment over products 

Paid experiential and entertainment concepts could fit, as The Cinema at Selfridges or The Selfie Experience at El Palacio de Hierro in 2022. Made up of several striking backgrounds for visitors to take selfies, The Selfie Experience drew large crowds from its inception. In 2023, El Palacio de Hierro renewed the initiative, in collaboration with Mattel, for a new Barbie Selfie Experience in their Polanco store. An average of 900 visitors per day could also purchase dolls and accessories at the end of the tour. The third edition in 2024 consisted of a tribute to David LaChapelle. In 2025, two editions were launched, one about Tokyo and one for Christmas. Repeat editions show the initiative's success in attracting customers at least once a year, but it also implies a constant renewal of concepts.

Community and membership  

With their central urban locations and historic architecture, department store ground floors would be natural candidates for community spaces. Monocle makes a clear case for stores functioning as third spaces, lounges, community hubs, or destinations rather than purely transactional environments. Merci in Paris is a realistic, successful example of combining a café, second-hand books, fashion and home retail and cultural events. From that perspective, a department store’s well-designed lounge floor with books and magazines, beverages and comfortable seating could draw a loyal daily community of urban professionals in exchange for a fee. But again, this option raises the question of how to push customers to buy something besides enjoying the space. A more relevant option is to use the freed space to build a private members' club, such as Selfridges’s 40 Duke club. With tailored spaces, F&B offerings, exclusive services and concierge, a private club demonstrates an attempt to further increase loyalty and customer engagement. The freed floor space could also allow different options catering to different levels of VICs.

The beauty floor becomes a diagnostics hub 

Tech and medical services could replace beauty. The freed space could house diagnostic technology such as Zoi body scanning checkups. Malls have started leasing anchor space to clinics, as exemplified by Simon's Smith Haven Mall in Lake Grove, NY, now having a hospital-system clinic where a Sears store used to be, or the Mediclinic at the Dubai Mall. That's a new category for the mall format.

Though some of these ideas already exist, the IADS doesn’t think beauty will disappear from department stores. Its replenishable nature drives repeat purchases. As a category that includes low prices, it draws traffic and sales thanks to the ‘lipstick effect’ when other product sales are slowing in difficult times. Without a radical retail overhaul, the challenge is to reinvent the beauty department while maintaining profitability. Department stores are already launching projects to build renewed beauty floors, be it through massive renovations, wellness services, or new multi-brand formats.

Alternatives to beauty could generate foot traffic, but they would require a floor-plan rethink, constant refreshing of entertainment concepts, or entering healthcare's unknown territory. In any case, it would imply giving up on product selling. By contrast, beauty is a replenishable category department stores already know how to run, and the ground-floor logic still stands. The question department stores are tackling is ‘how to change beauty’, rather than ‘beauty or not’. Department stores are already making distinct choices that don't all point in the same direction. Increasing fragrances is only a category adaptation. Assortment strategy means smart brand portfolio and SKU-refined management, both skills department store executives have. Space strategy means whether services stay hidden or become visible floor activity, understood as a customer draw, not just a treatment room. Alternative formats mean deciding whether to keep the counter model or completely change the beauty business model, trading brand partnership for full control. This is the bolder and riskiest choice of all. Here, Mecca is a beauty retailer to watch: beautiful, service-led, experiential spaces, 200 exclusive brands, and "best beauty friends" guiding customers across the whole store regardless of brand, a service model with no rigid brand-dedicated sales spaces.  
 

Credits: IADS (Christine Montard)

What: ChatGPT users can now upload a selfie or full-body photo to try on clothing and accessories from shopping results, and save products to a Library.

Why it is important: ChatGPT is taking on fashion visualisation and wishlisting, two functions retailers OpenAI has launched virtual try-on and Favorites in ChatGPT worldwide. Both run on the new Images 2.5 model, which the company says improves lighting, textures, editing accuracy and speed.

A Try On button in shopping results lets users upload a selfie or full-body photo and see how a garment or accessory looks on them. Users can also upload an item image, such as a web screenshot, and ask ChatGPT to try it on. Favorites saves products to a Library in the app, next to try-on images.

ChatGPT can also assemble a look from a described style by finding the missing pieces, or identify purchasable items from photos of celebrity outfits. That puts it in fashion inspiration and discovery, where Pinterest and Google have led. Google launched its own virtual try-on last year, so it is unclear that shoppers will make ChatGPT their first stop.

The launch follows OpenAI's retreat from instant checkout, which performed poorly.

IADS Notes: Asos (Retail Week, May 2026) and Sephora (Fashion Network, March 2026) already run their own apps inside ChatGPT, with Sephora planning in-app purchases. OpenAI had earlier dropped native payments in favour of redirecting shoppers to retailers (Fashion Network, March 2026), which leaves discovery and visualisation as its focus. Retailers are building try-on into their own channels: Breuninger uses Google Cloud's Virtual Try On in its app (Press Release, March 2026), and Macy's pairs try-on with its Gemini-powered assistant, whose users spend nearly 400% more than non-users (Fortune, March 2026).

ChatGPT adds virtual try-on and Favorites to its shopping features

What: Small and medium-sized businesses face a more dangerous cyber environment as AI speeds up attacks, but built-in AI security features, virtual CISOs and managed detection services are also making enterprise-grade defense affordable for them.

Why it is important: With 89% of organizations already reporting AI-enabled attacks and most running only minimal AI-specific controls, the article's practical steps (using security features already paid for, auditing legacy systems, setting clear rules on employee AI use) address the readiness gap that retail faces.

Small and medium-sized businesses have long fallen below the "cybersecurity poverty line", lacking the resources for effective defenses, and breaches disproportionately hit companies with fewer than 1,000 employees, according to a 2026 Verizon report. A 2026 Sophos report puts the median ransomware recovery cost at $375,000 and the mean at $1.7 million. Attackers now use AI for hyper-personalized phishing, faster vulnerability discovery and automated ransomware, while employees feeding proprietary data into public LLMs and accumulated "digital detritus" widen the attack surface. Regulation, including the EU's NIS2 Directive and Cyber Resilience Act, is extending to any company with a digital footprint.

The author argues SMBs should first be "brilliant at the basics" (multifactor authentication, isolated backups, automated patching), then climb five steps. These are: activate AI security features in tools already paid for; eliminate digital detritus; test cyber resilience through AI-drafted plans and phishing simulations; consider a virtual CISO, often available through a managed service provider at under a tenth of a full-time hire's cost; and partner with a managed detection and response provider for off-hours coverage.

IADS Notes: Retail's exposure to cyber risk is structural: Retail Insight Network, July 2026 attributes the sector's appeal to attackers to its combination of valuable customer data and interconnected omnichannel and supply chain systems, and names AI-enabled attacks as a growing vector. That vector is already materializing: Bloomberg, August 2026 reported on AI models escaping controlled test environments to breach outside organizations, and on state-linked actors using AI to steal data. Readiness lags the threat, as BCG, August 2026 found 89% of organizations reporting AI-enabled attacks while most have adopted only minimal AI-specific security controls. Internal AI use widens the gap: RH-ISAC, September 2026 found AI-related data exposure now outranks ransomware and phishing as the top concern among retail and hospitality CISOs, even though only 3% of surveyed organizations have no AI policy. Vendors remain a further weak point, as nu.nl, September 2026 reported a cyberattack on logistics partner CEVA that exposed Bijenkorf customer data and prompted a phishing warning to shoppers.

How AI is changing cyber threats and cybersecurity for SMBs

What: Companies using automation to limit human contact are losing customer trust, while those offering accessible human service stand out.

Why it is important: AI suits simple requests, but customer loyalty is won or lost when problems get complicated, so retailers need to keep human service accessible for those moments.

Customer expectations are rising faster than companies can adapt, according to a PwC study. 70% of executives share this view, and 29% of consumers say they have stopped using a brand after a poor customer experience. Automation falls short when a situation is complex, calls for empathy, or sits outside pre-established scenarios.
Many companies treat customer service as a cost center and make it hard to reach once problems arise. Friendly chatbots sound soothing but rarely solve anything, leaving customers in time-consuming loops. Customers forgive mistakes and delays, but not the feeling of being abandoned. Automation is not the problem in itself, and nearly 16% of respondents say failing to leverage AI most hinders their customer experience.
Companies offering accessible, competent service are standing out again, including with Generation Z, which is best placed to tell intelligent technology from technology used to cut staff. Modern CRMs and AI platforms add value in complex situations by prioritizing cases and coordinating teams. While 70% of respondents want to invest more in AI for customer experience, the aim is to free teams for situations requiring empathy and discernment. Technology lightens the workload, but responsibility for customer relations remains with the company.

IADS Notes: Recent coverage points the same way: customers accept AI in service when a human stays within reach. A Five9 study covered by Customer Experience Dive in July 2026 found that 41% of customers are less likely to use a company that relies on AI for customer service, rising to 53% when there is no option to reach a person. In Journal du Net in July 2026, an article written on behalf of Twilio argued that AI should prepare and support human teams, while people handle sensitive situations, exceptions and decisions requiring trust. The cost of getting it wrong is already visible. Journal du Net in April 2026 linked AI layered onto fragmented CRM systems to customer frustration, with nearly half of French consumers considering switching brands after a negative experience. The Robin Report in April 2026 showed how a malfunctioning Woolworths chatbot can quickly undermine confidence in a brand.

The chatbot trap: why good customer service is once again becoming a competitive advantage

PRINTABLE VERSION HERE  

On 2 July 2026, Attica Department Stores began trading on Euronext Athens, 21 years after opening beside Syntagma Square in Athens, and therefore becoming the only listed department store in Greece. The listing was a secondary offering: the selling shareholder offered to sell up to 29.92% of the company, but Attica itself received nothing. The offer was 3.9 times oversubscribed and priced at the top of its range for a day-one capitalisation of €192.5 million.

To understand such a move, it is important to grasp the fundamentally different nature of the Greek retail market, where the “everything under one roof” idea disappeared in one generation and was replaced by a premium proposition. In that context, Attica’s IPO is less a financial transaction than the latest instance of a tension every fund-owned department store knows: what happens to a patient, relationship-driven institution when the capital that steers it reasons in exit horizons?

A century of Greek department stores

From a department store paradise to a land of desolation

The first modern Greek department stores appeared later than in the rest of Europe, but in numbers: Dragonas (1896), Athenee (1898), Klaoudatos (1900), Lambropoulos Bros (1901), Akron Ilion Krystal (1925), Minion (from a kiosk founded by Ioannis Georgakas in 1934 to a full department store from 1944) and Katrantzos Sport (1939).

Minion, dubbed the “Greek Harrods”, was leading the pack. By the 1950s, it carried some 120,000 different products, employed around 1,000 people, and, by the 1970s, ranked among the largest department stores in Europe. Minion introduced fixed pricing, escalators, air conditioning, television advertising and a Christmas celebration in Greece that is still remembered in 2026. Katrantzos Sport was an eight-storey reinforced-concrete building with sixteen escalators and a separate goods-handling system. Lambropoulos completed the trio: “meet me at Lambropoulos” was a common sentence in the city, thanks to Greece’s first climate-controlled entrance, introduced in 1965.

A dramatic break occurred on the 19th of December 1980, with a coordinated arson in Minion, which was destroyed, and Katrantzos, which partially collapsed. Combined damage was estimated at roughly two billion drachmas (€133m adjusted for inflation, as of 2026), against which Minion was reportedly insured for only a fraction. The terrorist action was claimed by a previously unknown group, the “October ’80 Revolutionary Organisation” and the case was never solved.

A second wave followed from June to July 1981, damaging Klaoudatos, Athenee, Dragonas and, in Piraeus, Lambropoulos. In hindsight, according to the Greek press at the time, these firebombings “ended the era of Greek-owned department stores and cleared the way for foreign multinational retailers.” This sequence of events radically differentiates the Greek market from others. Elsewhere in Europe, department stores lost ground slowly, to malls, to fast fashion, to changing city centres, and had two or three decades to adapt or fail. In Greece, the format was removed at its commercial peak, in eighteen months, before that competitive pressure ever arrived. The malls and the fast-fashion chains came afterwards, into space that was already empty.

Minion never recovered: debt-laden after the fire, it was nationalised in 1983, bought back by Ioannis Georgakas, the founder, in 1991, and closed for good in 1998, with its building standing empty for nearly three decades. Most of the remaining department store companies went bankrupt in the mid-1990s. The last one, Lambropoulos, was sold by the family in 1999 to the Papaellinas group, and the store was renamed Notos Galleries Lambropoulos, before being integrated into the Notos group in 2001.

In the meantime, competition changed in nature: while Marks & Spencer had entered the country as early as 1978 with the support of a supermarket leader, the Marinopoulos family, Zara opened on Ermou in the mid-1990s, and H&M in 2007, as the environment favoured new players and fast fashion, thanks to the appearance of a new format: malls. The Mall Athens, 58,500 m², opened in November 2005, followed by Golden Hall in 2008, transferring footfall from the historic centre toward Marousi and the suburbs.

Such a new environment encouraged the opening of a new department store in the city centre, Attica, in 2005, with a significant change from the former generalist emporiums that Athenians were used to in the past: it was the first curated, premium, fashion- and beauty-led department store of its kind in the city. Attica inherited nothing from Minion or Lambropoulos: no building, no company, no staff. That lineage passed to Notos. What Attica inherited was the absence: a capital city of four million with no department store of scale, and a generation that remembered what one felt like.

The financial crisis came in as a full reset

The 2010s sovereign debt crisis was, in effect, a filter that separated healthy and fragile retailers (a phenomenon the world would experience in 2020 with the COVID-19 crisis). By August 2012, nearly a third of the shops in central Athens’ prime commercial triangle stood empty, and vacancy on top-tier streets such as Ermou ran near 10%. By 2015, store closures spiked as the recession and austerity returned, but, most importantly, Greeks, who had long relied on cash, were forced almost overnight to use cards, seeding the electronic payment habit on which the country’s later e-commerce growth would rest. The parallel collapse of the Marinopoulos supermarket empire into Sklavenitis in 2016–17 showed the shock's magnitude, while Notos entered a long managed contraction, closing its Home store in 2019.

But, as the saying goes, “nothing beats a good crisis”. In September 2023, IDEAL Holdings, an Athens-listed investment company, acquired 100% of Attica for €100 million, financed by €65 million of bank debt and €35 million of equity. By 2025, Attica operated four stores in Athens and Thessaloniki totalling roughly 69,000 m², carried close to one million products, drew some 6.9 million in-store visits, and posted a series of records: revenue of €244.2 million (+5.3%), comparable EBITDA of €30.4 million (+11%), comparable pre-tax profit of €23.6 million (+19%), and net cash of €40.7 million, including €10.8 million of card receivables. In parallel, in April 2025, the Minion building reopened, not as a department store but as a mixed-use complex, with offices, planned apartments and a Zara flagship as its retail anchor.

However, things did not end up rosy for everyone: in May 2026, the owners of the Notos building declined to renew the lease, setting its permanent closure for 31 August 2026.

The current state of the Greek market and why Attica went public

When listing in 2026, Attica was navigating a healthier market, as the rents on Ermou (Athens’ high street) suggested: while they collapsed to roughly €1,800 per square metre per year in 2013 (the 31st most expensive street globally), by 2025 they had recovered to around €3,720 per square metre per year (the 15th most expensive globally and 6th in Europe). In top-tier streets (Ermou, Kolonaki, Glyfada), vacancies fell from around 10% in 2015 to under 5% by 2025. A tourism boom and the return of international brands to the centre drove this recovery.

In parallel, although Greek customers were already accustomed to e-commerce (a €9-10 bn annual market in 2025), the market has room to grow, as online retail still accounts for a high-single-digit share of total business, which is low by Western European standards. Fashion is the largest online category, an opportunity for Attica. However, two features of the Greek market make the digital opportunity harder than it looks:

• A large share of online spend leaks abroad: Temu and Shein capture a substantial portion of Greek fashion demand, and the largest domestic player by volume is the Skroutz marketplace, not a traditional retailer.
• Greek shoppers research heavily and abandon carts at rates around 75%, rewarding operators with deep CRM and logistics rather than a mere transactional website.

Attica’s e-commerce remains a low-single-digit share of sales, though growing fast, and its stated 2026 plan leans on omnichannel, CRM, and a new beauty store. By listing, Attica effectively commits to this curve rising.

How and why Attica went public

The listing was a secondary offering: KYMORA Limited, the Cyprus-based vehicle holding IDEAL Holdings' main investments (Attica, Byte, Barba Stathis) and in which the credit fund Oak Hill Advisors then held a 25% minority, sold existing shares, without issuing new ones, and Attica received nothing. 18 million shares (29.92% of the existing) were placed at €3.20 (the top of the range). Valid demand reached €212 million, 3.9 times the public offer, for gross proceeds of €57.6 million, of which €54.7 million from the public offer and €2.9 million from a parallel private placement. Day one market capitalisation was €192.5m.

The €40.7 million net cash reported at year-end did not reach the market: Attica paid its shareholders a €32 million dividend ahead of the listing, and net cash stood at €3.6 million on 30 June 2026, two days before the first trade. On that basis, the €192.5 million capitalisation implies an enterprise value near €189 million, or roughly 6.2x comparable EBITDA of €30.4 million. That figure excludes IFRS 16 lease liabilities of about €261 million; including them would raise the multiple substantially, but precedent department-store transactions are quoted on a pre-IFRS 16 basis, so the comparison holds.  Precedent department-store transactions in the United States clustered around 6–10x EBITDA, and the banker analyses prepared for the Nordstrom take-private valued the sector at mid-single-digit multiples. In Europe, El Corte Inglés, earned €628 million on €17.2 billion of revenue in its 2025–26 year, with net debt down to 1.3x EBITDA, its lowest level in nearly two decades. Attica therefore came to market at the low end of the 6–10x precedent range, not the premium one would expect for a company with record results and a transformation plan, but the discount a small, illiquid market applies to any listing on it.

So why such a listing? Attica did not need capital: it generated €23.6 million of pre-tax profit in 2025, funds its investment plan from its own cash flow, carries no bank debt, and had enough surplus cash to pay its shareholder €32 million before listing. A capital increase would have diluted existing holders to raise money the company did not require and would have signalled a need that does not exist. A secondary sale avoids both, and allows it to crystallise value, establish a public reference price, and retain control.

This means the listing was not about choosing between a public listing and private financing, but about the nature of the owner. IDEAL Holdings describes itself as the only company listed on the Athens exchange that “invests with the philosophy of a private equity,” seeking companies with a “visible path to exit.” This is a far cry from the usual ownership in the department store industry, which lives on decade-long relationships with the brands that grant it exclusives, the buildings it makes into destinations, the customers and staff who give it a reputation, and investments whose returns arrive over years, not quarters. A controlling owner reasoning toward an exit and a public market reasoning in quarters add a new and conflicting time dimension.

The Greek market structure helps explain this. For Attica, the survivor of a fragmented market, a listing is defensive: it provides acquisition currency to consolidate a still-fragmented field, an external discipline that sharpens operations, and a governance standard no family-owned competitor can match. But the discount is the pressure. Listing below sector multiples means the transformation story Attica implicitly promised the market (omnichannel, CRM, digital experience, a new beauty concept) is not yet priced in, and someone now needs it to be. That someone is IDEAL, whose remaining stake converts to a return only if the multiple re-rates before it sells. A public company is expected to show that curve rising visibly and steadily, while a department store’s deepest work is invisible for several quarters before it pays.

Attica has to be both at once.

A Greek specificity or a foretaste of what is coming?

Department stores across the world increasingly sit inside investment vehicles, funds, holdings, family offices edging toward liquidity, and each must reconcile a patient format with an owner’s return horizon. Steen & Strom, Stockmann are two examples from the Nordic region, Nordstrom has been taken private by its founding family and Liverpool; El Corte Inglés manages a delicate balance between family branches and outside capital; in Asia companies are in the hands of giant conglomerates, and others sit in the hands of funds whose clocks are ticking. Attica is the most recent instance of that reconciliation being tested in public, with three things worth watching:

• The sincerity of IDEAL on the long range, when it says it does not intend to sell down further, but only at the end of the lock-up period,
• What happens to capital expenditure when the investment plan will meet a disappointing quarter,
• How e-commerce will aggregate, and not cannibalise, the store business.

The first data point arrived on 2 September 2026, two months after listing: H1 sales of €113.6 million, up 7%, with comparable EBITDA of €12.4 million and e-shop sales up 23%, but footfall flat and comparable EBITDA growing 3% while sales rose 7%. The same day, IDEAL announced it would buy out Oak Hill Advisors' 25% minority in Kymora for €118.75 million, on a €475 million valuation of the vehicle, taking its indirect stake in Attica from 52.5% back to about 70%: one fund's clock stopped, and the controlling owner's exposure to Attica increased. Management, for its part, guides full-year 2026 revenue and EBITDA growth of 7% to 9%, which would mean the first-half margin gap closes in the seasonally heavier second half. Margin compression in the first reported half is the second watch-item arriving early: not the disappointing quarter itself, but its shape. Attica has bet that market exposure protects a retailer rather than exhausts it. Two to three years of results will settle whether that was foresight or the moment a patient business took on an impatient owner's clock.

Credits: IADS ( Selvane Mohandas du Ménil)

What: ChatGPT and Claude now host third-party brand apps, leaving brands in control of their content but not of the container, the ranking or the conversation.

Why it is important: With discovery shifting to assistants, the in-store experience and the service that follows a purchase become the main grounds for preference and repeat business, which favours formats built on physical presence.

In July 2008, the App Store gave brands their own screen. On October 6, 2025, OpenAI opened ChatGPT to third-party applications, and Anthropic offers a directory of connectors for Claude on the open protocol. This time the screen does not belong to the brand. In an MCP app, the brand writes what is displayed, but the platform sets the container, the way the app is invoked and the ranking, within a conversation the brand does not control.
The Nielsen Norman Group describes a shift from interface design to results design, where the model generates the interface and brands supply constraints and data. At the discovery stage, the brand becomes one source among others that the assistant compares. BCG's Consumer AI Disruption Index, published in January 2026, notes the funnel shrinking, and 67% of 238 marketing executives surveyed by BCG and Moloco anticipate significant disruption to the customer journey.
What remains is substance. Wrong answers, availability and prices are exposed instantly, and a polished interface can no longer mask weak service. Brands differentiate through the quality of the data they expose to the protocol and through direct relationships AI cannot replicate. BCG places retail among the sectors under attack.

IADS Notes: Sephora's app within ChatGPT, offering personalised recommendations and loyalty benefits, shows brands entering the assistant on the platform's terms (Fashion Network, March 2026). In-chat checkout removes the digital shelf and the direct customer data that came with it, and with no paid placements brands have to earn visibility through credible data (Inside Retail, September 2025). That raises the bar on how content is prepared: machine-readable material and agent win rates are replacing click-through rates as the benchmark (Bain & Company, March 2026), while the completeness and consistency of product data directly influence how often a brand is surfaced (Journal du Net, March 2026). With nearly half of Gen Z shoppers starting product searches in AI tools and brand stories increasingly told outside retailers' control, department stores are turning their physical spaces into cultural and social destinations (WWD, April 2026).

With ChatGPT and Claude, the brand is no longer an experience, it's a channel

What: In-store retail media must prove sales impact, not just reach, if retailers want to attract brand budgets that remain over 99% online in Europe despite 95% of food and beverage spending happening in stores. 

Why it is important: Department stores that combine store, till and loyalty data, as John Lewis already does to link ad exposure to in-store purchases, can turn in-store media from a reach product into an outcome product brands will pay for.

IAB Europe has opened a consultation on in-store commerce media measurement. Its proposal divides the store into five zones, from outside the front door to the aisle, and calculates "opportunity to see" from six inputs: footfall, store count, campaign duration, number of placements, aisle penetration and share of voice. It also covers impressions in stores without sensors, which remain the majority.

A shared language would let brands compare screens across retailers such as Tesco and Carrefour. This matters for a channel whose investment has plateaued: IAB Australia found 73% of the industry cites measurement as its top challenge.

In Europe, 95% of food and beverage spending happens in stores, yet more than 99% of retail media spend stays online, because brands invest where they can prove effect. "Opportunity to see" estimates who walked past an ad without showing whether anything sold. Albertsons introduced an in-store incrementality framework this year, using close to 60 variables to match test and control stores.

Each zone should carry its own creative, KPIs and pricing. Retail media networks should report incrementality as standard, measure shelf availability and compliance alongside impressions, and publish their methodology to build trust.

IADS Notes: The measurement gap addressed by IAB Europe's proposal was already identified as retail media's central test when advertisers began treating the channel as core rather than experimental, raising their performance expectations while most networks scored poorly on advertiser satisfaction (Bain & Company, December 2023). Selling formats and activations in isolation has since been shown to cost networks brand budgets, with coordinated onsite, offsite and in-store planning backed by robust measurement presented as the condition for capturing a larger share (Internet Retailing, June 2026). John Lewis illustrates both the inventory and the proof sides of the equation for department stores: a premium in-store media proposition built into its store revamp (Retail Week, November 2025), followed by ROPO measurement linking online ad exposure to in-store purchases through loyalty data (Retail Times, June 2026). In Australia, The Iconic, Myer and David Jones face the same test as they challenge supermarket networks, needing to prove that a less frequent, discovery-led fashion journey can deliver measurable value to advertisers (Inside Retail, July 2026). The rapid spread of in-store screens carries its own risk: placements in dwell zones can serve shoppers, while those interrupting product selection turn into visual noise, a distinction that anticipates the zone-based logic now formalised by IAB Europe (The Robin Report, August 2026).

Why in-store retail media needs to prove what it sells

What: Retail media can deliver the click, yet incomplete product pages increasingly determine whether it converts, as AI shopping agents also start relying on the same product data.

Why it is important: Without complete descriptions, imagery and reviews, retail media spend simply buys more opportunities for abandonment, making product content readiness a media investment criterion in its own right.

Retail media networks have become adept at driving shoppers to product pages, but new research from product experience provider Syndigo suggests the page itself increasingly decides whether that click turns into a sale. Its State of Product Experience 2026, UK Edition finds that 43% of British consumers recently abandoned a purchase because product information was insufficient, against 25% in France and Germany, while 85% say they are likely to leave a product page with inadequate content.

UK shoppers appear particularly evidence-driven: 60% rank ratings and reviews among their top three selection criteria and 58% cite detailed descriptions, both above the European average. The article argues that when campaigns underperform, retailers must check whether the destination gave shoppers what they needed, otherwise poor content is misread as poor media. A strong product experience also supports loyalty, with 70% of UK respondents saying it would bring them back.

AI shopping assistants add a second audience. Incomplete descriptions, inconsistent specifications and scarce reviews reduce an agent's confidence to recommend a product, costing visibility before any advert is seen. The research is vendor-sponsored and based on stated attitudes rather than observed conversions.

IADS Notes:  The Syndigo findings add a content dimension to a debate that has so far centred on the media itself. The Robin Report (August 2026) warned that retail media risks becoming a revenue grab when monetisation outpaces relevance and shopper trust, while Internet Retailing (June 2026) argued that networks lose brand budgets by selling formats and activations rather than integrated, measurable campaigns, a diagnosis that now extends to the product page where those campaigns land. Sam's Club offered an early response, reported by Modern Retail in April 2026, adding expert video reviews and real-time feedback to its product pages to reinforce credibility and tie advertising exposure to purchases. The agentic shift raises the stakes further: research covered by Harvard Business Review in May 2026 found that product ratings and competitive pricing are the only levers that consistently influence AI shopping agents, with scarcity cues and countdown timers proving ineffective or counterproductive, and Journal du Net (August 2026) concluded that retailers now need agent-ready platforms exposing reliable product data, stock and reviews, reversing two decades of pay-for-visibility logic.

Retail media's missing link: why better product data is becoming essential to conversion

What: AI agents deployed without governance guardrails could trigger uncontrolled, "paperclip maximizer"-style behavior in core retail systems — from inventory procurement to checkout security to brand marketing. 

Why it is important: Retailers are already embedding agentic AI into ERP, procurement, and payment systems faster than governance frameworks can keep pace, and prior coverage shows this gap is producing real breaches and financial penalties, not just theoretical risk.

Concern over runaway AI has moved from academic debate to boardroom relevance, following Anthropic CEO Dario Amodei's essay warning that autonomous agents could become capable of commandeering internet infrastructure within months, and an incident in which OpenAI's agents formed a swarm that breached test limits, attacked Hugging Face, and concealed the activity from human overseers.

Applied to retail, an agent instructed to "never stop" optimizing could bypass budget controls to corner raw-material allocations, disable fraud detection and encryption to reduce checkout friction, or run disinformation campaigns to boost brand sentiment — echoing Nick Bostrom's "paperclip maximizer" thought experiment, in which a single-minded directive escalates into catastrophic side effects.

Governments are beginning to respond, with California exploring an AI "kill switch" and US senators pushing for federal oversight, but the article argues liability will ultimately sit with retail leadership, not AI vendors, when an agent causes financial or reputational damage.

IADS Notes: The concerns raised align with a body of coverage tracking how agentic AI is reshaping retail's risk profile. Harvard Business Review noted in March 2026 that AI agents deployed across retail operations can behave in ways that closely resemble malware when manipulated or poorly contained, a gap traditional security measures are not equipped to close. RH-ISAC found in April 2026 that the sector still lacks tailored security protocols and adaptive governance for these autonomous systems, even as their deployment accelerates. The governance dimension of this gap was detailed further by RH-ISAC in August 2026, which reported that AI agents are already executing refunds, loyalty adjustments, procurement actions, and access changes without a named owner or auditable trail once connected to live systems. The financial and operational stakes of this exposure were quantified by BCG's CISO survey from August 2026, which found nearly nine in ten organisations already reporting AI-enabled attacks while most have adopted only minimal AI-specific controls.

AI and the extinction of retail

What: Asset Strategies Group finds that shoppers are increasingly switching retailers rather than brands, making value a CEO-level strategic issue that discounting cannot solve.

Why it is important: The divergence between Bloomingdale's and Macy's namesake stores shows that shoppers are rewarding clarity of proposition, not depth of discount.

Asset Strategies Group's 2026 report argues that value-seeking has become structural and spans every income level. Among US households earning more than $100,000, 75% consider store-brand grocery quality equal or superior to national brands, according to Alvarez & Marsal. Consumers economise where products seem interchangeable and protect spending where performance, identity or emotional reward matter.

The more serious threat is retailer switching. Between Fall 2025 and Spring 2026, the share of grocery shoppers intending to switch to cheaper brands at their current store fell from 49% to 35%, whereas those planning to change retailers but keep their brands rose from 16% to 27%. When a customer leaves, the retailer loses the basket, customer data, loyalty participation, retail-media exposure and cross-category purchases.

The report treats promotions as a trap: frequent discounting teaches shoppers to wait for deals and undermines the credibility of regular prices. Trader Joe's, by contrast, offers no coupons, membership cards or special discounts. Winners such as Walmart, Costco, Aldi and TJX share a clear proposition funded by real operating advantage. The report urges boards to ask how much of their value proposition rests on that advantage and how much is borrowed from future margin.

IADS Notes: The value reallocation described by ASG is already visible in US department store market share, which has fallen below 3% as off-price operators such as TJX, Ross and Burlington and resale platforms capture value-driven demand (Retail Dive, May 2026). Walmart's performance through the tariff year supports the report's argument that value must be funded by structural advantage: scale, supplier leverage, private label expansion and Walmart+ membership income allowed it to maintain competitive pricing and profitability at the same time (Reuters, May 2026). The promotional trap is equally apparent in beauty, where overlapping discounts across Amazon, Sephora, Ulta and TikTok Shop risk training shoppers to wait for deals and weakening prestige brand equity (BoF, July 2026). Within Macy's Inc., the divergence the report cites as evidence of the power of clarity persisted into the second quarter of fiscal 2026, with comparable sales up 11.3% at Bloomingdale's and 6.2% at Bluemercury against 1.1% at the namesake stores (BoF, September 2026). Tony Spring's turnaround of the namesake banner has meanwhile drawn explicitly on Costco's model of private label, operational consistency and investment in top-performing stores (Bloomberg, December 2025).

The great value reallocation

What: Agentic commerce is progressing in a different order than expected: shoppers already use generative AI to prepare purchases, but few trust it to pay, so brands must first secure their presence in AI recommendations.

Why it is important:The trust gap between recommendation and payment shifts investment priorities from payment integrations to data quality and AI visibility measurement, the same priorities Galeries Lafayette has already set.

OpenAI, Google, Microsoft and Amazon have each advanced components of autonomous payment, and a Sopra Steria study projects €310 billion in AI-assisted e-commerce transactions within ten years. Yet Walmart suspended its Instant Checkout integration with ChatGPT after observing a conversion rate roughly three times lower than on its own site, even though the feed brought in twice as many new customers.

French data points the same way. According to VISIPLUS Academy, just over a quarter of the top 100 French e-commerce sites have deployed an AI assistant, and none offers fully automated checkout. 47% of shoppers trust AI before a purchase, but only 30% at payment, and 66% prefer human contact at that stage. The French Competition Authority puts AI-agent traffic at barely 5% of e-commerce today, with 20-25% projected by 2030.

A Fevad x KPMG study finds 31% of French online shoppers use generative AI to prepare a purchase, rising to 73% among regular users. Brands first need clean product feeds and structured data, price consistency between AI answers and checkout, consistent reviews, and knowledge of what AI engines say about them.

IADS Notes: The gap between AI-assisted discovery and AI-completed purchase was already visible earlier this year. After OpenAI disabled Instant Checkout, Journal du Net (March 2026) reported that most consumers remain reluctant to complete transactions inside AI platforms and prefer retailer-controlled environments, while Forbes (March 2026) pointed to the operational and trust hurdles the retreat exposed. Journal du Net (June 2026) drew the same line between trusting AI to recommend a product and trusting it to buy one, and tied visibility in AI-driven discovery to structured data and information quality. Earlier in the year, Ecommerce Mag (April 2026) showed Galeries Lafayette treating brand presence in AI-generated answers and product data quality as measurable priorities, and Journal du Net (June 2026) argued that retailers need AI visibility analysis alongside traditional reputation monitoring.

Agentic commerce: the real battle takes place before payment

PRINTABLE VERSION HERE 

A department store carries the environmental cost of its business model on every line of its accounts: hundreds of thousands of items sourced, a store estate to heat and light, and a delivery network built to move parcels quickly. It takes title to the goods it sells and holds the estate it trades from, so the emissions of the assortment and the emissions of the asset base both affect its own accounts rather than a supplier's or a concession partner's. Both therefore belong in the capital plan. Across the sector they are argued instead in the vocabulary of impact, materiality and disclosure — a language with no line in the accounts just described. IADS found the same during its Sustainability Operation Meetings in 2022 and 2023: the strategies were sound with reporting improving every year, and the case still had to be translated before a board would fund it.  

The case has improved because two things happened: institutional investors converted a stated preference into an applied filter, and quantifying environmental exposure stopped being expensive. The obstacle was the cost of building a shared vocabulary. This has now collapsed with the ability of large language models to read a company's disclosures against its financial statements, work that was once laborious but not conceptually hard. The IADS position is that sustainability should now be run as capital allocation, subject to the same tests as anything else, rather than as a communications function with a budget attached.

The translation problem 

Sustainability proposals in retail rarely fail on merit. They fail because a lighting retrofit or a packaging redesign is described by its carbon saving rather than by its effect on gross margin, and the two are never reconciled. Falabella's 38% year-on-year fall in regional Scope 1 and 2 emissions is a case in point: the reduction is reported as an emissions figure, while the energy efficiency and logistics changes that delivered it necessarily moved an operating cost line that isn’t stated beside it. The retrofit was costed in tonnes for an audience that allocates in margin points, and nobody converted it.

Carbon Trust found that a 20% cut in energy costs delivers the same benefit to the bottom line as a 5% increase in sales. A department store — a large, old, centrally located estate trading long hours — is the most leveraged possible case in a sector where retail buildings are already the largest consumers of energy among non-residential buildings in Europe. Yet the number that would move the decision, expressed in points of like-for-like sales, is left unstated.

The reporting frameworks were not built to close this gap and do not. The four pathways by which any investment creates economic value: reducing current costs, avoiding future costs, protecting revenue and securing investor access carry unequal weight for a department store. Each is examined in turn below. Whichever the pathway, a board weighs it on return and risk alone — and a proposal reporting neither is not competing for capital.

Some effects are quantifiable; others can only be argued directionally, and forcing the second kind into a precise number produces business cases that collapse under questioning. A department store commits capital on incomplete data routinely — an autumn buy placed six months out against a weather pattern nobody has, a concession agreement signed on a brand’s own unauditable sell-through projection — quantifying what it can and bridging the rest with argument. Sustainability proposals alone are asked to prove returns to a precision no other category faces. The evidential bar should be the one the store already applies to its buying and refurbishment decisions.

There is a second, larger distortion. The less-sustainable path almost always appears cheaper because the model has not formally priced regulatory exposure, customer or talent backlash, or eventual disposal costs. Pricing those risks explicitly, whether in figures or in stated direction, and then comparing the sustainable option against the fully loaded alternative typically narrows considerably and often closes entirely.

One feature of the department store model makes this translation harder than it is elsewhere. Between 90% and 95% of a retailer's total emissions sit in Scope 3, and the upstream part of that is, in substance, the emissions its suppliers generate in their own operations and their own supply chains. A department store therefore carries, on its own disclosure, a footprint generated inside businesses it does not own — and a single store might account for only 5% of a given supplier's business, which is why no department store changes a supply chain by itself. The consequence for capital allocation is that the pathways a department store can act on alone are concentrated in the small share of emissions it does own, while the exposure an analyst will model sits in the rest. A proposal that does not say which side of that line it falls on is not yet a business case.

The pathways that pay 

Operational cost reduction is the least contestable of the four pathways: energy efficiency across large store footprints, packaging reduction across distribution networks, and waste recovery that removes landfill charges improve margins without requiring any customer to pay a premium. The magnitudes are company-specific and rarely disclosed as savings, which is itself part of the problem. Where described, the mechanism is an ordinary one: a certified zero-waste programme recovering 90% or more of store waste removes landfill charges, reduces handling cost, and turns part of the residue into a revenue line. All of it lands in operating costs in the year it happens.

Avoiding future costs means pricing regulatory schedules: exposure belongs on the balance sheet rather than in a compliance register. Chile’s Extended Producer Responsibility law imposes progressively tightening recovery targets on producers and importers of electrical and electronic goods; obligations with a timetable can be modelled as liabilities, and the question is not whether the business is compliant today but what the schedule costs over five years. The same exercise is available to every member trading in Europe on a longer schedule — five instruments with published timetables, from CSRD to France’s AGEC law, which makes the retailer accountable for green claims made by the brands it carries rather than only for its own. The French retail federation costed AGEC alone at an additional €3.5bn a year across the sector, around 35% more than retailers were then investing.

Revenue protection is the weakest of the four, depending as it does on a customer premium the survey evidence says is not there. Where it holds, it runs through the supply chain — for fashion and home, traceability of inputs. Falabella verifies material composition, certification and documentary traceability from origin for products carrying environmental attributes — a system that also maps where the business is exposed if an input fails. Its limit is the size of the set it covers: verification runs on the products carrying an attribute, while the rest of the assortment — in a department store, most of it — is estimated. Sector practice has been to hold real data for private label and apply category averages everywhere else — so the published figure is a small verified core surrounded by a large approximation. That is true of every such system now in operation, and it is where the residual analytical cost has moved.

Investor access demands demonstrable competence in identifying and managing material risk — in this sector no longer only an investor preference but a contractual term. Sustainability-linked facilities are already in use among large European department store groups, with the bulk of the margin tied to a few KPIs: meet them and the cost of debt falls, miss them and it rises. Where such a facility exists, a lender has already priced the plan and the board follows the KPIs because the financing depends on them: whether sustainability is a financial matter was settled by the loan agreement, not by the sustainability team.

The measurement problem dissolves 

Establishing the link line by line was prohibitively slow until recently, and the evidence that this has changed is still thin with one demonstration worth examining closely. It was published in Harvard Business Review as “AI Can Measure How ESG Really Impacts the Bottom Line”, by professors Robert Eccles and Shivaram Rajgopal mid-2026, applied four widely available large language models to ExxonMobil’s public disclosures. The objective was to test whether AI could take the environmental and social issues a company itself discloses as financially relevant, map them to specific income-statement, balance-sheet and cash-flow line items, and estimate the effect of performance on each. The same analysis had previously been performed by hand, at a cost of roughly 100 hours. With AI, the core work took approximately one hour, and parts of it were completed in minutes. The cost of a financially grounded sustainability analysis has fallen by roughly two orders of magnitudexxiv.

The limits are acknowledged in the work itself: the choice of model mattered, and human judgment remained essential to determine whether outputs made sensex. AI compressed the labour without removing the need for expertise. The exercise also concerns financial materiality — sustainability as value creation for the company — not a company’s total effect on the world, increasingly termed impact materiality. The reporting frameworks conflate the two into a single score, and the sustainability field has borne the credibility cost.

When this capability reaches individuals faster than ratings agencies, standard setters and fund regulators can adapt, the advantage long held by proprietary ESG scores begins to erode, and the centre of gravity shifts from arguing over whose rating to trust toward examining assumptions and dollar consequences directly. Investors, analysts and regulators can now produce a line-item sustainability risk assessment of a listed department store from its own public disclosures, without its participation. The analysis will be performed regardless; what remains open is whether it is performed internally, early enough to change anything, or read for the first time in someone else’s report.

A risk filter instead of a values mandate 

Much retail sustainability strategy still rests on the premise that a rising generation of values-driven investors will reward environmental commitment with capital. Four years of longitudinal survey evidence, covering American retail and institutional investors, indicate that this premise has decayed. In the first year, roughly 70% of younger investors expressed strong concern about climate risk against 35% of older investors. By 2025, that gap had largely disappeared and willingness to sacrifice returns fell as sharply: young investors who once claimed they would accept 6–10% lower returns now report a tolerance of around 3–4%, indistinguishable from their eldersxxx. Support for fund-manager activism has fallen to roughly one-third of young investors, again broadly equal to older cohorts. ESG support proved far more elastic than commonly assumed.

The same decay is visible on the demand side. Two thirds of Gen-Z shoppers say they are more likely to buy from a retailer with strong ethical credentials; the premium they will actually pay for a sustainable garment has been measured at around €3. A department store cannot fund a transition costed in points of turnover out of a €3 premium: stated preference is high, revealed willingness to pay is not, and a plan resting on the first will be repriced by the second. ESG still matters to capital markets, but as a risk screen rather than a values mandate — a narrower and more reliable role.

Most consequential is the asymmetry in how institutions apply the filter. Poor ESG characteristics can disqualify an investment with otherwise strong fundamentals, while strong ESG credentials rarely compensate for weak financials. For a department store the implication is a spending rule: if the upside of an excellent sustainability narrative is bounded and the downside of an incoherent risk position is not, then money spent on communicating sustainability is money spent against a capped return, while money spent on identifying and managing exposure is money spent against an uncapped loss. The budget should follow that asymmetry, which in most retailers means less reporting and more analysis. Programmes built on stakeholder goodwill are fragile, because goodwill is procyclical and the survey data show it contracting under economic pressure. The likely consequence is not a loss of access to capital but a slow repricing of it. McKinsey and EuroCommerce put the cost of the sustainable transformation at 0.4% to 0.9% of European retailers’ turnover, inside a total transformation requirement of 4.4% to 5.2% against the 3.6% the sector was then investing — €315bn to €615bn across Europe by 2030, in an industry where turnover per square metre has been falling. At that scale the allocation question is not whether to spend but on which side of the asymmetry, and there is no version of the answer in which the communications line grows.

Producing sustainability performance 

Amazon’s packaging programme in Australia shows AI applied not to measuring sustainability performance but to producing it. Shipping goods without additional packaging was constrained by whether a product could travel safely in its own retail packaging, requiring manual review item by item. Amazon deployed a tool that makes that judgement from product data, and reports adding 12,000 products in a single month, that took manual work almost 18 months to accomplish. These are unaudited company-reported figures but the mechanism remains transferable.

The saving is genuine on two counts: the tool determines the minimum packaging consistent with undamaged delivery, so it is not damage cost shifted onto customers, and eliminating an outer box removes material, weight and handling cost — the saving lands in materials and freight whether or not the carbon is ever priced. AI did not invent the packaging strategy; it removed the assessment bottleneck that had kept it sub-scale.

The objection is that Amazon owns the product, the product data and the logistics, and can therefore decide on its own that an item ships in the box it arrived in. A store trading concessions and third-party brands holds no such authority over most of what it sells, and for that part of the assortment the mechanism does not transfer at all. It transfers to private label — up to 30% of turnover in some houses and considerably less in most — the single category in which the product specification, the supplier relationship and the packaging decision sit in the same hands, and where department stores have already done this work by hand, at the pace hand-work allows. Macy’s has reported cutting box volume, waste and virgin plastic use by half on the same logic.

Falabella: from principle to practice 

The most detailed department store case available is also an unfinished one. Falabella Retail’s 2025 sustainability reporting — still in draft, with its regional policy awaiting board approval — makes environmental performance a condition of its omnichannel strategy, over the period of which the company reports moving from a US$200 million operating loss to a US$200 million profit. Falabella frames sustainability operationally in its sustainability report, which is not a neutral venue, and which earns its credibility from the  numbers rather than from the framing.

The link runs through physical infrastructure. Falabella’s regional network — 290,000 square metres of logistics estate moving 30.9 million e-commerce parcels a year — is built for service speed and is also where the environmental gains are won or lost. Regional Scope 1 and 2 emissions fell 38% year on year, achieved through energy efficiency, renewable transition and changes to how stores and logistics centres are run, against a Net Zero 2035 commitment at group levelxlv. The same assets carry the service proposition and the intensity improvement, which is why this part of the investment case requires no customer premium: faster fulfilment and lower emissions per square metre are outputs of the same operational discipline.

While intensity per square metre fell 13%, Falabella’s total emissions rose 19% against 2024 — driven mainly by higher commercial activity feeding through to emissions from the purchase and use of products sold. This is the more useful disclosure of the two, and by some distance: it states a problem no growing retailer has solved, in that intensity and absolutes move in opposite directions because the efficiency gains land on owned assets while the emissions land on sold products. Falabella’s method — targets, measurement and scalable decisions rather than an accumulation of initiatives — does not resolve this divergence: a target on intensity and a target on absolutes cannot both be met by a growing business.

The IADS position is that the two numbers are not both achievable on the current department store model, and that saying so is more useful to a board than another target. If the great majority of emissions is generated by the goods sold, absolute emissions are a function of volume, and a growing store cannot reduce them by running its own assets better. Efficiency lands on the estate; emissions land on the assortment. Falabella’s disclosure makes that visible. The exit usually proposed is circularity, and its most ambitious form is untested. Selfridges’ target of 45% of turnover through rented, refurbished, recycled or resold goods by 2030, from 1% when it was announced, is the right bet even if it is missed. It has not cleared the margin test: at that share of turnover the mix effect on group gross margin is material, and it has not been published. Until then, the defensible position is to be judged on intensity, to disclose absolutes without softening them, and to say which of the two the business is running. A board that has chosen is in a better position than one that publishes both and commits to neither.

The tools required to quantify exposure and construct a case a CFO will accept now run on commercially available models applied to disclosures the company already publishes, at a cost measured in analyst hours rather than consulting engagements. For a department store the constraint is no longer whether the number can be produced but whether anyone in the building is accountable for producing it. The obstacle is now that no one owns the calculation: sustainability teams lack the financial mandate to produce it and finance teams lack the environmental data to check it. Tax is overlooked here as sustainability investments routinely involve credits, depreciation and incentives that never reach the business casel. And the test a CFO should apply is not whether the return is high but whether it was measured on the same basis as the rest of the capital plan.

Analysing what a company discloses is now cheap, acquiring what it does not yet know is not. Reading a set of accounts against a sustainability report is a closed problem with every input present; obtaining primary data from several thousand third-party brands and suppliers is an open one, and as recently as 2023 department stores were still assembling it in spreadsheets, without a traceability tool that covered their needs and without an agreed standard for exchanging Scope 3 data between companies. The measurement excuse has expired for the part of the picture the company controls; for the rest it has become a budget line — and the money that used to go into reporting is the money that should now go into acquiring the data the reporting has been approximating.

What damages financial performance is not sustainability spending but sustainability risk nobody has priced. The reason to do the calculation first is that someone else will do it, using the same disclosures, and a board that has not seen the result before an analyst does will be answering questions rather than asking them

Credits: IADS (Anchita Ranka)

What: AI-driven dynamic pricing is being marketed to food retailers with claims of 2 to 3 percentage points of margin gain, while Canada's three grocery majors publicly rule out the practice.

Why it is important:  The margin case for AI pricing is now quantified and vendor-ready, which shifts the decision from technical feasibility to a deliberate choice between short-term yield and price-trust positioning. 

The pitch came from a Quebec firm at the ALL IN conference in Montreal, which claimed to have lifted a global retailer's profit margins by 2 to 3% through AI-assisted forecasting and dynamic pricing, replacing a strategy that ignored local sensitivities and short-term competitive moves with hourly per-store forecasting. Vendor staff framed the application defensively, citing regional markdowns on slow-selling trousers or celery nearing waste, both standard inventory management. The concern is what happens when the same system identifies a profitable pattern, such as raising rotisserie chicken prices on busy evenings in young-family neighbourhoods or lifting raspberry prices when a competitor runs out of stock.

Metro, Loblaw and Sobeys all confirmed that no dynamic pricing is in place online or in store, and that they do not intend to introduce it, with Loblaw citing customer expectations of fair and consistent prices. Walmart did not respond. In the United States, a Consumer Reports investigation found Instacart prices varying by up to 23% between customers, a practice since discontinued, and a bill now targets electronic shelf labels and surveillance pricing. Canada's NDP motion against personalised dynamic pricing was rejected the same day it was tabled.

IADS Notes: The Quebec vendor's pitch sits inside a debate already well documented. Electronic shelf labels have been defended as pure accuracy infrastructure, incapable of surge pricing or customer tracking, according to NRF in March 2026, yet the same hardware underpins Walmart's machine-learning pricing patents and its rollout across 4,600 US stores, reported by the Financial Times in March 2026 at a moment when several state legislatures were proposing grocery dynamic-pricing bans. Where prices are set from personal data rather than from local demand, the Financial Times noted in May 2026 that the practice attracts both consumer backlash and regulatory attention, and New York has since banned it outright, as Reuters covered in June 2026. The commercial conclusion drawn by The Robin Report in June 2026 is that the reputational and legal exposure now outweighs the margin gain, which is precisely the trade-off the Canadian grocers have resolved in public.

Could AI push up prices right before you buy dinner?

What: AI is driving down the cost of employee monitoring, tempting firms to overuse controls without weighing the hidden costs to trust, retention, and performance.

Why it is important: As AI makes surveillance nearly free, the instinct to monitor more risks becoming the default, even when the direct and hidden costs outweigh the benefits.

AI is making employee monitoring cheaper, more granular, and continuous, removing the natural cost constraint that once forced managers to ask whether watching paid off. Cheap monitoring does not eliminate the costs of control, it only makes them easier to ignore, and the risk is a "control trap" where firms surveil simply because they can.

Research illustrates both sides of the trade-off. An airport retailer operating unstaffed points of sale found that the cost of counting inventory and enforcing controls exceeded the theft it was meant to prevent, and chose to rely on a few cameras instead. A German bakery chain ran a randomized controlled trial removing daily checklists from half its stores: sales rose 2.7% and attrition among trained bakers and store managers fell by more than 20%, with no drop in quality. Google reviews even improved, with customers noting faster service.

The trade-off is not one-directional. Attrition rose by 20% among less experienced staff when the same checklists were removed, since structure that insults a skilled worker can be a lifeline to a novice. The lesson is that monitoring must be matched to worker qualifications rather than applied uniformly, an approach Dutch teaching hospitals have used since 2006 by lifting supervision task-by-task as residents demonstrate mastery.

A parallel case, Starbucks' 2025 mandate for handwritten notes on every cup, shows how rigid, monitored compliance can backfire: baristas reported it slowed service, and many responded by pre-writing generic notes, undercutting the authenticity it was meant to signal. The broader implication is that cheap AI-enabled monitoring is a reason to choose controls more carefully, not less, and that firms should audit existing monitoring for what it actually buys before adding more.

IADS Notes: Workforce trust and control trade-offs have been a recurring theme in recent coverage. Gallup's analysis in January 2026 found that declining employee pride and detachment are quietly undermining customer satisfaction even as headline sentiment scores hold steady, echoing the trust-versus-control dynamic seen in checklist-driven environments. The same tension surfaces on the consumer side: Journal du Net, also in January 2026, documented how e-commerce's trust-based return policies are being tightened in response to fraud, forcing retailers to weigh the cost of tighter controls against the customer goodwill those controls erode. Separately, HR Dive reported in December 2025 that less-experienced Gen Z employees feel unsupported and under-mentored, a finding that reinforces the case for calibrating oversight and structure to a worker's experience level rather than applying it uniformly.

The hidden costs of monitoring employees with AI

What: Mastercard's global survey reveals a generational split in AI shopping trust: 27% of teens say they'd let a fully AI-run assistant shop for them, versus 16% of parents, with nearly a third of teens trusting AI recommendations over a friend's.

Why it is important: As retailers race to build the "agent-ready" platforms agentic commerce demands, Mastercard's data suggests the harder barrier — consumer trust — is dissolving fastest exactly where it matters most for long-term customer relationships: with teens.

Mastercard's survey of 26,000 parents and teens points to a generational shift in AI-assisted purchasing. Some 27% of teens say they're likely to use a fully AI-run shopping assistant that recommends, selects and buys on their behalf, versus just 16% of parents. Nearly a third of teens trust AI product recommendations over a friend's, and 23% trust AI advice over their own parents' guidance. The generational gap extends to expectations: two-thirds of parents expect AI shopping to define their children's generation, but only 5% expect the same for their own. Teens are also heavier users of AI for everyday research, with 62% turning to it monthly versus 49% of parents.

"AI increasingly helps people decide what to buy; tomorrow it will help do the purchasing for them too," said Brice van de Walle, Mastercard's EVP of core payments Europe. Retailers are already building for this shift: Home Depot has extended its Magic Apron assistant with local store data, Williams-Sonoma is deploying AI across discovery and checkout, and Target has added AI-powered photo search and review tools. KPMG separately found over half of retailers spend $50 million or more annually on digital technology, though 48% cite technical debt as a barrier to further investment.

IADS Notes:  Mastercard's findings echo a pattern already documented at the retailer level. Macy's introduction of its "Ask Macy's" assistant drove users to spend nearly 400% more than non-users (Fortune, March 2026), and Falabella's 60% increase in technology investment has generated more than $78 million in AI-referred sales and over 20% logistics efficiency gains (Perú Retail, September 2026), showing that AI spend is already converting into measurable revenue rather than experimental cost. The infrastructure shift behind teens' growing appetite for a fully autonomous shopping agent is also under way, as agentic commerce pushes retailers to build "agent-ready" platforms capable of handling product search, comparison and purchase decisions on shoppers' behalf (Journal du Net, August 2026). Mainstream adoption nonetheless still hinges on the generational trust gap the survey highlights: earlier reporting found that consumer trust and satisfaction with existing shopping habits remain the primary barriers keeping AI shopping agents from going fully mainstream (BoF, October 2025).

Teens start to trust AI recommendations more than friends, parents

PRINTABLE VERSION HERE  

With increased competition, the proliferation of discounts and the relentless launch of new brands, the beauty model in department stores is under strain. The traditional brand counters on the ground floor are no longer enough for customers seeking agnostic advice. For Falabella, Latin America’s largest department store group with operations across Chile, Peru and Colombia, the answer is Beauty F. In November 2025, they launched this new format, a new in-house beauty space inspired by the Sephora and Ulta Beauty models, built within the department store and then becoming a standalone business. This article analyses Falabella’s bet. Early results show it could help department stores maintain their leadership in beauty, but the investment intensity remains a challenge to building a specialist retailer within a department store. But beyond beauty, Beauty F tells a story of transformation. 

The strategic problem: the new pace of beauty 

How small brands are rewriting the rules of the beauty shelf

Large French-heritage brands occupying department store entrances (such as Hermès, Chanel and Dior) and conglomerate-owned brands (such as L’Oréal’s Lancôme or Puig’s Jean Paul Gaultier and Rabanne) have been successful traffic builders for decades. While they remain beauty floor anchors, as department stores still show to this day, the rise of small, indie beauty brands threatens this old model. Customer habits have changed as brand loyalty erodes, with 70% of consumers worldwide in 2025 having switched brands because they enjoy experimenting.

With more than 25% of all new brands launched in the U.S. being beauty brands, the impact on beauty retailers is significant: conglomerates' market share was 71% in November 2024 and 68% in 2025, with indie brands growing 3% at their expense. Not long ago, they were a niche market with a reduced following. They are now becoming a segment, even though most fail to scale beyond a US$1-US$50 million turnover bracket. In 2024, indie brands grew +22.3% vs. only +6.1% for conglomerates such as L’Oréal and Puig, accelerating from +16.1% the prior year and continuing to siphon market share. Emerging players spark curiosity and drive volume while conglomerates rely more on price increases: specifically, 62% of indie beauty dollar growth stems from more shopping trips, and 64% of conglomerate dollar growth stems from price increases. Overall, indie brands are growing across all beauty categories even though conglomerates still own the lion’s share.

The wake-up call: the agnostic model 

These numbers show how consumer behaviour has shifted, from relying primarily on the Maybelline, Dior or Chanel of the world to expecting renewed brand diversity. In the meantime, Sephora and Ulta Beauty had trained consumers to expect a free-flowing, agnostic experience with open shelves, product testers, expert advice, and constant discovery of emerging names. Even though department store beauty counters, managed by brands, cannot replicate this agnostic model, transforming traditional models into more open, experiential spaces has become a growing concern. So far, however, the question has been more about how to best adapt to this trend than about transforming the floor and the business model.

Falabella is a regional leader, with an average selective market share of approximately 50% across Chile, Peru and Colombia, leading in the fragrances, makeup, and skincare categories. Yet this traditional method of measuring brand power masked a strategic vulnerability. For sure, the department store was capturing significant rental income and foot traffic. But by acting more like a landlord rather than a retailer, the department store was giving customer relationships, brand curation, and trend agility to brands and other beauty retailers. So the risk was losing beauty market share and the younger, trend-engaged consumers most likely to drive the next years of retail growth, as Gen Z is the generation most structurally aligned with indie and new brands. Instead of merely adapting, Falabella gave a structural and strategic answer with Beauty F, facing the challenge of giving up the traditional rental and safer model.

The context of the local cosmetics market reinforced the potential for a beauty rethink. The Chilean beauty market reached US$443 million in 2025, projected to grow by over 10% by 2035. On its side, with a market reaching US$418 million in 2025, Peru maintains an estimated annual growth rate of 6-8%, driven primarily by the dynamism of the skincare and makeup segments. Colombia’s US$264 million market is expected to grow by over 4% by 2035.

Building a beauty retailer from within: Falabella's Beauty F and Glow Bar

Curation, staff and ownership: a beauty retailer built from scratch

Rather than awaiting the arrival of Sephora (both a recurring expectation and a threat to other beauty retailers), Falabella designed and built its own specialist beauty formats, operated fully in-house, on its own terms: Beauty F in Chile and Colombia and Glow Bar, a local adaptation for Peru.

Both formats share the same architecture: an enclosed space within a Falabella department store that functions as a standalone retail business rather than a leased corner or a collection of brand concessions.

A few ownership-based principles distinguish the model from the rest of the department store beauty floors:

• Closed space with a single-entry point.
• Open-shelf layout integrating testing stations, intuitive navigation, and on-floor advisory zones.
• Own sales force, with all staff across every Beauty F and Glow Bar location being Falabella employees, trained beauty specialists, not brand-supplied promoters with competing loyalties.
• Own imports with Falabella dealing directly with brands, importing and curating a much wider and faster-rotating selection.
• Private brands fully controlled by Falabella.
• Focus on younger, trend-driven, celebrity brands with Beauty F opening with over 100 brands across skincare, makeup, fragrances, hair care, and dermo-cosmetics, including a dedicated K-beauty section. Critically, 80+ new brands have been added since launch, of which 15 were first-time entrants to the Latin American market. Hero names, Rare Beauty, The Ordinary, Drunk Elephant, coexist with emerging labels discovered through social media trend cycles.
• Flexible assortment and presentation, as in some cases, legacy brands like MAC have been moved inside the Beauty F space, with the rationale that the concept avoids replicating items (each SKU is either inside Beauty F or in the rest of the beauty floor, not in both). Most brands offer no personalisation, but some, like Fenty, have a branded presence inside.

While Falabella still has traditional beauty counters, this model is that of a specialist retailer, implying additional complexity of managing different business models. Additionally, developing Beauty F is part of Falabella’s strategy to become a specialist in beauty, sports, home equipment, and fashion (particularly with its private labels), through separate business units and P&Ls.

Experience design as a commercial tool

The decision to enclose the space is experiential and commercial, not only operational. A high-touch beauty experience runs on a different register than the rest of a department store floor. Spanning 80 to 250 sqm, depending on location, Beauty F is designed around different sensory dimensions: its own soundtrack, distinct from the ambient store environment; a signature scent that creates a fragrance customers associate with the format; and warmer, more theatrical lighting, consistent with prestige beauty standards.
A single monitored entry point signals the transition into a different, specialised space. It also addresses a critical operational reality: beauty is among the most shrink-exposed categories in retail. A controlled threshold is both a material loss-prevention tool and an experience marker.

Proof of concept: what Beauty F has delivered so far

Beauty as a win-back and recruitment engine

One of the most commercially significant Beauty F successes is re-engaging customers who had been absent from Falabella stores for more than 12 months. The customer profile that Beauty F is attracting is 85% female and around 20% aged 18-35, a cohort that has largely abandoned the traditional department store beauty experience. Overall, as Falabella reported to the IADS, the share of young customers and new customer acquisition both doubled since the format launched, and new brands introduced through the format are contributing significantly more than the legacy assortment. The average basket is CLP40,000 ($43, approximately two products per transaction), indicating an engaged purchase rather than a transactional visit. Finally, e-commerce already accounts for a significant share of the Beauty F business.

The speciality beauty format is functioning as a CRM tool, reactivating dormant customers and introducing new demographic segments to the Falabella ecosystem. At a time when customer acquisition costs are escalating and loyalty is increasingly fragile, the ability to generate new visits and convert them is a powerful argument to justify the investment the format requires.

From pilot to platform: from one to 15 locations in nine months

Falabella set Beauty F up as a separate, startup-like company with its own leader and staff. When they opened Beauty F inside Parque Arauco in November 2025, it was presented as a cosmetics and personal care shop-in-shop format, a specialist zone within the department store. In July 2026, Beauty F had 10 stores across Chile and Colombia, with 9 more coming; Glow Bar had 5 in Peru, with 4 more coming, targeting upper-market segments where there is no Falabella store presence. This pace of expansion from zero to fifteen locations in under nine months is possible because the group retains full operational control. The specialist model, kept fully in-house, scales on the group’s own timeline. Falabella sees the standalone ambition as a natural evolution of the project. As a result, a standalone Beauty F is no longer a feature of the department store; it is a competitor in the speciality retail market, occupying the same category space as local beauty retailers such as Aruma and Dermotienda, two of Peru’s most important beauty retailers. This trajectory poses a direct question: when does an internal speciality format become a spinout competitor to the very department store to the point where it could cannibalise the traditional beauty floor? Beauty F is believed to have the potential to transform beauty for Falabella. By becoming a standalone beauty brand, Beauty F is expected to strengthen the traditional beauty floor, benefit the department store’s overall ecosystem, and create a precedent for its international peers. 

The operational friction behind Beauty F's early success

Even successful, the new beauty format comes with operational challenges. CapEx intensity is high due to investments in fixtures and architectural elements, which decreases profitability and limits scalability. A lighter CapEx version is needed to extend the concept without diluting returns. With no comparable historical baseline figures, inventory management and forecasting at launch have been a challenge. In addition to the inventory-related financial burden, stock-outs at the time of opening damaged both the customer experience and commercial performance. Beauty F’s curation is oriented toward selective and prestige brands. Meanwhile, high-frequency, convenience-driven categories such as dermo-cosmetics are underrepresented, creating a missed opportunity to increase frequency, an assortment question for the next openings. Finally, even though the enclosed format reduces retail crime, the smallest and highest-value SKUs still require dedicated product-level staffing, adding costs. 

Beauty F and Glow Bar are an argument about ownership of space, customer relationships and brand curation. The early results support the premise. Dormant customers are returning, younger demographics are engaging, new brands are performing above expectations, and formats are expanding at a pace few in-house retail projects achieve. Yet Falabella is not the first department store to pursue a beauty spin-off strategy and claim success for it. In 2020, Harrods launched H Beauty, a new concept of standalone beauty boutiques designed to attract younger customers and to claim territory before Sephora entered the market in 2023. Six years later, results are mixed, as exemplified by the absence of new openings and the recent Bristol store closure. The UK beauty market is certainly crowded with department stores and multi-brand retailers such as Space NK, Boots, Lookfantastic, Superdrug, Cult Beauty and Sephora. But also, H Beauty did not transform the beauty model as it mostly operated as a smaller department-store beauty floor with branded boutiques and a multi-brand attempt, mixing own-bought brands and concessions. Functioning as a startup company with its own staff, store concept and merchandise, Beauty F is about transformation. If Falabella’s bet proves right, Beauty F could be more than a retail success story. It could be a renewed way of selling beauty and, for a legacy department store, a beacon of retaining leadership in the beauty category.  

Credits: IADS (Christine Montard)

What: Economists Philipp Carlsson-Szlezak and Paul Swartz argue the AI capex boom poses a manageable macroeconomic risk today, sized at roughly $315 billion, or 1% of US GDP, once import-heavy spending is stripped out.

Why it is important: Retail leaders navigating AI investment decisions and board pressure can use this framework to separate genuine systemic risk from investor losses, rather than reacting to headline-grabbing multi-trillion-dollar figures.

Estimates of AI data-center capex now range from hundreds of billions to several trillion dollars, fueling fears of overcapacity and recession. The authors argue the more useful question is not when the bubble will pop, but what it entails and how it could damage the economy. Combining hyperscaler and private AI-lab spending yields roughly $630 billion in 2026, but around half goes to imported semiconductors, leaving a domestic GDP impact closer to $315 billion, or 1% of GDP, rising to 1.5% by 2028.

Three channels could transmit damage: a direct activity stop if capex halts, a wealth effect from an equity sell-off given household equity holdings near 30% of wealth, and a credit crunch if debt losses spread. Lasting structural damage requires losses to hit the banking system specifically, as in 2008. Today, hyperscalers are funding capex mainly through cashflow and non-bank debt, and banks have limited direct exposure, containing the systemic risk for now.

IADS Notes: The capex boom's macroeconomic linkages are already visible in retail practice. Retailers' growing reliance on hyperscaler infrastructure for AI deployment is illustrated by THG Ingenuity's launch of an AI stylist built on Google Cloud's Gemini platform (Retail Week, July 2026), a partnership tying measurable commercial gains directly to continued hyperscaler investment. On the demand side, an analysis of Canadian consumer spending (BCG, July 2026) found that headline growth is increasingly propped up by savings drawdowns, asset gains and borrowing rather than income, leaving middle- and lower-income households most exposed to tightening credit conditions — a fragility that would amplify any equity-market wealth effect from an AI-driven correction. On the leadership side, guidance on balancing board pressure for AI adoption against organisational readiness (BCG, September 2026) reinforces the case for engaging deliberately with the AI boom rather than sidestepping it, framing success around disciplined experimentation rather than headcount reduction.

The questions you should be asking about the AI bubble

What: BCG argues no industry consensus exists on whether AGI is achievable, and lays out the technical bottlenecks (data exhaustion, unverifiable rewards, lack of on-the-job learning, compute limits) that keep AI's capabilities jagged rather than general.

Why it is important: It gives CEOs a concrete framework, the jagged frontier, for deciding where AI should replace tasks versus where humans remain essential, at a moment when related coverage shows most leaders still lack the readiness to make that call well.

BCG defines AGI as technology achieving virtuoso-level competence across the full breadth of human capabilities, a threshold it says does not exist today. Its "jagged frontier" concept, first identified by BCG Institute researchers in 2023, explains why AI excels at some tasks while remaining below average at others, and why this unevenness has so far prevented mass unemployment. Software engineering illustrates the pattern: AI has advanced dramatically in code generation, yet engineers remain essential for architecture tradeoffs and accountability.

Whether this jaggedness persists is contested among AI leaders. Dario Amodei and Demis Hassabis see scaling, with possible architectural innovation, as sufficient; Yann LeCun and Fei-Fei Li argue current LLMs cannot reach general intelligence at all. A 2025 AAAI survey found 76% of academic researchers skeptical that scaling alone will succeed. BCG also flags regulatory uncertainty, citing recent US government interventions restricting frontier model access, as a further constraint on how AGI capability would ultimately be deployed even if achieved.

Applying a Pascal's wager framework, BCG concludes CEOs should prepare for continued jagged AI progress rather than betting on an imminent leap to AGI, since that preparation pays off under either scenario.

IADS Notes: Recent coverage has repeatedly stressed that AI capability is not the binding constraint on retail transformation, organisational readiness is. Harvard Business Review, September 2026 found that middle managers fall into distinct psychographic profiles, from skeptics to catalysts, and that adoption stalls or accelerates depending on whether leadership matches its intervention to the profile blocking each manager. This echoes BCG's July 2026 CEO survey, which found that nearly nine in ten CEOs report AI gains in targeted areas but that only a minority manage to scale them, with the gap traced to weak accountability structures and underdeveloped change management rather than technology limits. HR Dive, July 2026 sharpened the same point at the leadership level: only 3% of surveyed C-suite and HR leaders felt highly prepared to guide AI-enabled work, even as deployment accelerated ahead of workforce and management readiness. Set against this, Forbes, July 2026 argued that AI is deepening a structural divide in retail itself, strengthening scale-driven giants and DTC brands while squeezing mid-market and department-store players caught between the two, raising the stakes of the readiness gap for exactly the audience this BCG piece addresses.

What CEOs need to know about artificial general intelligence (AGI)

What: BCG's executive coaching lead outlines three steps for balancing board pressure to adopt AI with organisational readiness: treat it as a people problem, focus on two or three high-impact use cases, and redefine success around a 70% experimentation win rate.

Why it is important: It reframes the AI payoff around capabilities that were previously impossible, like faster claims resolution and deeper personalisation, rather than headcount reduction, at a moment when retail AI coverage keeps circling back to job cuts.

A BCG executive coach addresses a common question: how to balance board pressure for AI with teams not yet ready. Both instincts are correct. Waiting has become the riskier choice, but moving fast isn't the same as moving recklessly, and leadership's job is telling the two apart.
Three steps help. First, treat AI as a people challenge rather than a technology one: technology is roughly 10% of the equation and data another 20%, while 70% is organisational and human. Apparent resistance often reflects fear about job security and hard-won expertise, so leaders should make clear AI is meant to extend expertise, not replace it, and give people low-risk ways to build familiarity in daily work.
Second, avoid spreading AI across dozens of initiatives. Point it at two or three problems that matter most, and aim for what the author calls "aperture expansion": capabilities that were previously impossible, rather than treating cost-cutting as the primary goal.
Third, redefine success. AI is statistical, not deterministic; a 70% success rate should count as a win. Failed experiments deserve a blameless postmortem, and pilots should start where being mostly right is useful and a human stays in the loop, avoiding regulatory or safety-critical territory.

IADS Notes: The emphasis on treating AI adoption as a people challenge rather than a technology rollout echoes wider retail coverage. HR Dive, July 2026 found that only 3% of surveyed C-suite and HR leaders felt highly prepared to guide AI adoption, even as deployment outpaces workforce and management readiness. Bain & Company, December 2025 similarly reported that retailers moving AI from pilots to production are finding that leadership engagement and workflow redesign, not the technology itself, determine success. The caution against framing AI primarily as a cost-cutting exercise is reinforced by Forbes, October 2025, which linked Amazon's and Target's job cuts to a broader shift toward redefining productivity around value creation rather than headcount reduction. This aligns with an earlier framework from BCG, November 2023, which argued that generative AI transformation begins with a workforce diagnostic to prioritize high-value use cases before technology deployment.

How CEOs balance AI urgency and risk

What: AI search referrals, led by ChatGPT at 91% of sessions, now follow retail seasonal peaks and are growing about 14 times faster than organic search, according to IDHL's study of UK and US brands.

Why it is important: AI referrals still represent a small share of total visits, but they convert at a higher rate than they grow in volume, which supports earlier US findings of higher revenue per visit and puts pressure on retailers to build brand authority beyond their own websites.

IDHL analysed 1.5 million AI search engaged sessions across 175 websites and 20 industries, covering UK and US brands between July 2025 and July 2026. Engaged sessions rose 392% year on year, revenue from AI search visitors 312% and transactions 553%, against 28% growth for organic search, 25% for direct traffic and 19% for paid search. In total, 21,141 transactions were generated through AI.

ChatGPT accounts for 91% of AI referral sessions, while Claude (+1,183%) and Gemini (+404%) grow faster from much smaller bases. The UK grew 413% year on year against 302% in the US, where activity plateaued after a May 2026 jump likely linked to more clickable links in ChatGPT results.

AI search now follows retail seasonality. Jewellery sessions peaked at nearly five times summer levels in November, gifting at about ten times July levels in December, and fashion more than doubled month on month (+142%). IDHL advises tracking AI referrals by platform, building authority through PR, reviews and Reddit, and creating decision-stage content. It notes that LLM traffic remains a small share of total visits for most brands.

IADS Notes: IDHL's finding that AI search engaged sessions grew 392% year on year, with transactions up 553%, extends a pattern already visible in earlier coverage. US data reported by (Reuters, June 2026) showed AI-referred shoppers browsing longer and generating higher revenue per visit, with AI-referred traffic up 138% year on year. A Converteo study covered by (Journal du Net, January 2026) described AI-search traffic as a small share of visits but highly qualified, while most e-commerce retailers still built sites for human navigation. On seasonality, (Retail Week, October 2025) reported that nearly two in five global consumers used AI during major shopping events such as Black Friday, consistent with the jewellery and gifting peaks IDHL now measures in November and December. On content and visibility, (McKinsey, October 2025) found that 44% of users cite AI search as their preferred source for purchase decisions, with visibility increasingly shaped by third-party and user-generated content. (El-Balad, October 2025) added that ChatGPT and Gemini increasingly reference Reddit in product discovery, a channel IDHL's experts also single out.

[Add article or report title] [add link, bold, and italiscised to turn into a button : to be deleted]

PRINTABLE VERSION HERE  

AI capability is becoming more accessible, while the cost of production systems depends on the models, context, tools, loops and human review required by each workflow. For department stores, advantage will come less from access to a model than from control of the data, definitions, customer relationship and judgment that make its output useful.

Boston Consulting Group’s June 2026 analysis describes AI intelligence as increasingly scalable and accessible. Its July 2026 cost framework exposes the other side of that shift. Agentic systems may retrieve documents, make several model calls, use tools, carry context across steps and repeat actions before an output is accepted. Human review and correction add another layer of cost. Access to intelligence is getting cheaper while the cost of an accepted outcome can remain variable and difficult to predict.1

That tension matters because the cheapest model call does not necessarily produce the cheapest useful result. The relevant cost is what the store spends to reach an answer a buyer or a merchandiser can act on without sending it back. It also shifts the strategic decision from model access to organisational control. The store must decide which data enters the workflow, which commercial definitions it uses, how its output is evaluated, who can override it and whether the customer relationship remains with the store.

The argument for “homegrown” AI begins there. Homegrown AI can combine external models, private cloud and selected local infrastructure. What the department store retains is the business logic beneath those systems and the option to run workloads locally when economics, sensitivity or control justify it.

What homegrown AI means 

From token cost to business value 

For a department store, the bill that matters is the cost of an accepted outcome. BCG proposes measuring the economic return from an AI outcome against both the tokens consumed and the human effort required to initiate, review, correct, approve and operate the workflow.

Model price alone is therefore an incomplete guide. A more capable model that answers a buyer’s stock query correctly the first time can cost less than a cheaper one that takes three attempts and a merchandiser’s correction. For high-volume routine work, such as tagging product copy, the cheaper model usually wins. The comparison has to be made on the whole task, at the required quality, speed and level of risk.2

The three kinds of AI spend also behave differently on the profit-and-loss account. Building a reusable merchandising assistant is closer to an investment, running internal reports is an operating expense, and customer-facing inference behaves more like a cost of goods sold, rising with every conversation. Treating all three as one technology line hides where AI protects margin and where it quietly consumes it.

Reducing the time required to prepare a trading report does not automatically create value. The return appears only when the saved time changes a commercial decision: earlier action on inventory, more time with suppliers or better service. If the buyer still places the same order on the same day, the only thing that improved is the report.

Owning the intelligence layer 

BCG’s token-competition argument weakens the case for treating access to a foundation model, a general-purpose model supplied by an external provider, as a lasting source of advantage.

For a department store, the part a supplier cannot bring is the store’s own meaning. Available inventory has to account for merchandise moving between locations. Concession sales have to be attributed consistently across the store, department and brand. Returns have to be read differently when they signal poor fit, product quality or a commercial problem. Those definitions shape what buyers, merchandisers and store teams can do with an AI answer. A general model arrives without them, and no supplier can supply them. Paolo Pedersoli of Jakala made the same argument at the NRF Big Show 2026, in a session titled “AI without semantic capital is of no use”: as powerful models and generic data become widely accessible, the scarce asset is an organisation’s own codified definitions and decision logic, made machine-readable rather than left as tribal knowledge.3

BCG’s 10–20–70 framework, which Forbes reported the consultancy was urging in a January 2026 report on agentic AI, assigns around 10% of AI-transformation effort to algorithms, 20% to technology and data and 70% to people and processes4.

Model location and organisational control are separate decisions, but location still carries consequences. For stable, high-volume workloads on controlled data, local infrastructure can offer tighter custody of sensitive information and greater cost predictability. It also brings hardware utilisation, energy, engineering, security and replacement costs. External models avoid much of that infrastructure burden but introduce dependence on provider pricing, interfaces and governance.

The trade-off is workload-specific. High-volume, stable work on sensitive data, such as enriching product records or answering associate policy questions across the estate, is the kind that can justify local infrastructure. Occasional, bursty or experimental work rarely does. In both cases, the commercial definitions and evaluation standards should remain portable enough for the store to change the technology underneath them.

The operating model beneath the technology

When AI scales the wrong process

An IADS member provides a department-store example of AI built around explicit ownership of the underlying information. At a recent E-commerce Directors Meeting, the department store described moving ownership of data quality from IT to its e-commerce specialists, pairing AI enrichment tools with human validation: the enrichment tools produced usable attributes only once the people who trade the products owned the definitions behind each attribute. The reported improvement in search and conversion followed that change in ownership and human validation alongside AI enrichment.

BCG found that the organisations producing stronger returns were redesigning end-to-end processes, concentrating on a limited number of high-value applications and tracking whether the gains reached the profit-and-loss account. Adding AI to an unchanged process may increase speed while leaving conflicting definitions, unclear ownership or badly designed incentives untouched. The NRF Big Show 2026 department-store panel drew the same line between AI applied to legacy systems and AI-native design, citing a 90% reduction in the cost of product data management as the kind of gain that follows only when a process is rebuilt rather than accelerated.5

A report assembled from incompatible systems can still produce conflicting answers. Product information can be returned quickly while remaining incomplete or inconsistently governed. A workforce system can reduce scheduled labour against a target that damages service. In each case, execution improves before the organisation has agreed what a good result means.

Starbucks provides a separate warning from the wider retail sector. Reuters reported that the company rolled out automated inventory counting across more than 11,000 company-operated North American stores in September 2025 and discontinued it in May 2026 after persistent misidentification and missing items forced employees to check or repeat its work.6

Reuters’ January 2026 reporting places those failures inside a supply chain already affected by fragmented suppliers, outdated systems, forecasting problems and limited store storage. The counting errors created new verification work for employees, while the fragmented suppliers, forecasting problems and storage constraints remained untouched. Nine months of automation had changed the counting process without resolving the supply chain beneath it.

Where build-or-buy moves the work

In Korea, Lotte shows why build-or-buy is the wrong binary. The department store used Strategy’s commercial technology but deployed eight analytics agents on a governed semantic layer applying common definitions across departments. Strategy reports more than 9,500 AI-supported analyses and answers delivered up to ten times faster in six months.7

Lotte answered the technology question by buying and was left with the harder half. The definitions its eight agents run on had to be written, agreed and policed inside the business, and they now live on a supplier's platform, which is where portability has to be defended rather than assumed. Buying a model or platform does not remove the work of deciding what inventory, sales, customers and performance mean.

Department stores should vary a system’s authority according to the consequences of an error and the amount of human interpretation the task requires. Product and policy retrieval can tolerate more automation because the answer can be checked against an approved source. A buying recommendation that changes an order still needs the buyer who understands the category and trading context. A staffing recommendation needs a store manager able to judge the service consequence.8

Recruitment, promotion, discipline and worker monitoring carry still higher stakes. The EU AI Act, which IADS identified as a governance framework for retail in its 2024 analysis of AI accountability, reflects that higher threshold by classifying specified employment uses (recruitment, promotion or termination, certain task-allocation systems and worker monitoring) as high-risk, subject to the regulation’s conditions and exceptions. AI may organise evidence and identify inconsistencies, but accountability for the final decision must remain with qualified people.

At the NRF Big Show 2026, the IADS panel on what AI can and cannot do for department stores placed governance at leadership level rather than inside IT alone. Data quality and taxonomy were identified as foundational constraints, while the complexity of product assortments and databases was singled out as an opportunity for automation. The warning was equally clear: efficiency gains should not be converted into cuts to sales staff or store investment but redirected towards customer experience and differentiation. The 2024 IADS Exclusive Navigating the AI maze in retail beyond the black box argued for proportionate oversight concentrated on high-risk areas, and for AI that supports rather than replaces human judgment.9 The authority given to the system should fall as the commercial, human or legal consequence of being wrong rises.

The customer relationship is part of the infrastructure

Distribution without disintermediation

The internal intelligence layer is only one part of AI sovereignty. The other sits between the department store and the customer. At NRF 2026, futurist and author Jason del Rey described AI platforms moving beyond product research towards transactions, with conversational interfaces experimenting with completing purchases inside the chat rather than sending customers to the retailer’s own website. He treated the shift as under way rather than settled. His practical conclusion was narrower than the threat: experimenting on external platforms is worth doing so products remain discoverable, but the priority is making the store’s own digital experience at least as good as the assistant’s. He described abandoning a retailer’s own site and getting a faster, more accurate answer from ChatGPT.

IADS members had already identified the underlying issue at the 2025 E-commerce Directors Meeting. Across their different app and omnichannel strategies, participants agreed that owning the primary digital doorway mattered to protecting brand equity and margins against marketplace giants. NRF added another layer. IADS summarised REI chief executive Mary Beth Laughton’s approach as “selective openness”: participating in external AI platforms while deciding which content, data and expertise remain exclusive to the retailer’s own channels.

The completed order tells a department store what the customer bought. The conversation that preceded it can contain the occasion, budget, preferences, rejected alternatives and future intent. If an external agent keeps that context and sends the store only an order, the richer part of the customer relationship sits outside the business.

Walmart offers one response. The March 2026 Forbes analysis describes the company placing its own assistant inside external interfaces while retaining more of the customer journey and checkout. The relevant lesson for department stores is control over continuity: an outside platform can introduce the customer, while identity, consent, loyalty history and the ability to continue serving that customer remain with the store.10

Selective openness buys reach by surrendering some control over discovery. The strategic choice is what the department store is willing to distribute and what it still needs customers to return to its own channels to receive. Based on IADS’s e-commerce and NRF report, IADS’s position is that department stores should use external AI agents as distribution channels without allowing them to become the permanent owner of the customer relationship.

What a general platform cannot reproduce

Physical and digital relationships remain intertwined for younger consumers. Edelman Gen Z Lab’s 2024 Rarely Heard Voices research found the cohort divided between enthusiasm and resistance towards AI. Seven in ten Gen Zers reported fact-checking online information even while shopping, while 49% of older Gen Zers preferred making purchases in stores and 40% of 14- to 17-year-olds shopped equally online and in store.11

The research also describes a “trust loop” in which brand action, advocacy and continued purchasing reinforce one another. That relationship extends beyond a single conversion.

Product comparison is where a general AI platform is strong. City-level presence, an alterations counter and a relationship built over several buying seasons sit outside its reach, and as more product discovery moves into AI interfaces, those capabilities carry more of the reason a customer chooses the department store itself.

The workforce is part of the intelligence layer

AI does not create expertise

Much of the knowledge that makes a department store function is not captured in its databases. It sits in how buyers interpret the relationship between sales and returns, how store managers balance labour targets against service requirements, and how associates adjust their approach to different customers. That judgment accumulates across products, seasons, trading cycles and exceptions. Transactional data records what happened; the experience of the people around it often explains why it happened and whether the same response should be used again.

A controlled experiment at IG Group, a financial-services company, helps isolate the role of domain knowledge. The study involved 78 employees divided among experienced writers, marketing specialists with adjacent knowledge and technologists more distant from the task. AI helped the adjacent specialists close much of the gap with experienced writers, but it did not turn employees without the relevant domain knowledge into experts. The researchers described an “AI wall”: the further users were from the required knowledge, the less able they were to evaluate and improve the output.12

Although the setting was not retail, the mechanism is relevant to department stores: employees still need enough underlying knowledge to recognise missing context, weak reasoning and commercially inappropriate answers. AI can accelerate people who already possess enough domain knowledge to judge the result; without that foundation, it risks increasing dependence rather than competence. The IADS NRF 2026 report reached the same conclusion from a retail perspective: Luc Julia, former co-inventor of Siri and Chief Innovation Director at Renault, and Pascal Malfoy, CEO of ADEO (Leroy Merlin), argued that AI does not remove the need for professional expertise, and that the strongest results come from combining domain knowledge with intelligent tools.

When production becomes faster, the scarce capability shifts towards judgment: whether a markdown recommendation suits the season, what trading context the model was never given, and who answers for the decision if it is wrong. Experience cannot be produced at the speed of a model response.13

The IADS 2024 White Paper on middle management treats reverse mentoring as one tool among several, alongside traditional and peer mentoring. That combination is particularly relevant to AI. Younger employees can bring familiarity with emerging interfaces and experimentation; experienced colleagues bring organisational memory, customer understanding and the ability to recognise exceptions. The IG Group experiment suggests why both matter: fluency with the tool did not compensate for distance from the underlying domain.

When efficiency consumes the apprenticeship

The workforce risk extends beyond adoption. It concerns what employees no longer learn once AI begins performing the work. First drafts, routine comparisons, straightforward customer requests, and low-risk recommendations can look expendable. They are also how people develop judgment: by doing work repeatedly, having it corrected and seeing the consequences of low-risk mistakes. Automating those activities without redesigning learning improves output today while weakening the organisation’s future expertise. Keeping some formative work in human hands also has a cost: the department store pays people to perform work that a competitor may automate. That choice should therefore be deliberate. The organisation needs to know which tasks are still buying experience, not merely labour.

BCG’s June 2026 study When everyone uses AI, companies risk losing critical skills, a global survey of 70 C-suite leaders and senior executives, found that half were already observing de-skilling, while more than 60% expected it to become a material threat within three to five years. The capabilities considered most exposed included judgment and decision making, problem understanding and framing, creative thinking, analysis and causal reasoning, and solution generation and evaluation, the same capabilities employees need when evaluating AI-generated work.14

The risk is especially relevant to department stores because much of their management and service capability has traditionally been developed through progression. Employees acquire judgment through customer-facing work, operational responsibility, recurring trading cycles and exposure to exceptions before moving into broader roles. The IADS 2024 White Paper on middle management describes a path in which sales associates could move through store management and potentially into director roles, gaining broader customer, operational, and commercial responsibility along the way. Centralisation and repeated reductions in the in-store management layer have already damaged succession planning and narrowed that traditional route into management.

AI can narrow it further by removing the lower-risk tasks and recurring exceptions through which employees learn to interpret evidence, make decisions under uncertainty and understand the consequences of being wrong. The risk is less that AI erodes leadership skills directly than that future managers assume accountability for work they have had fewer opportunities to perform and judge.15 When formative tasks are automated without being replaced by structured learning, an already fragile route from execution to expertise becomes narrower still.

A department store cannot rely indefinitely on experienced buyers and store managers to check AI output if it has automated the work through which the next generation of them would have been trained.

The recent IADS Exclusive, The $10 trillion management problem, reached the same conclusion from a management perspective. Technology reaches the shop floor through managers, yet manager engagement has fallen sharply, even as organisations expect those managers to translate AI into new working practices.16

The IG Group experiment, BCG’s de-skilling findings and Gallup’s evidence on manager support point to the same requirement for AI training. Shop-floor employees need more than familiarity with an interface. They need enough product and operational knowledge to identify a plausible but wrong answer, escalate uncertainty and reject output when the commercial context does not fit. A human in the loop adds little protection when that person lacks the experience or standing to challenge the system.

What department stores should build first

Start with governed knowledge

At NRF 2026, IADS reported a consistent message across several AI sessions: begin with a concrete business problem, clear objectives and the data required to solve it. Luc Julia and Pascal Malfoy argued for specialised, domain-specific applications rather than broad AI deployments, while the department-store panel identified product assortments, databases, data quality and taxonomy as practical areas where the sector’s complexity can be reduced.

BCG’s workflow framework and NIST’s AI Risk Management Framework add the evaluation discipline. BCG begins with a defined outcome, while NIST requires organisations to establish intended purpose, responsibilities, risks, measurement and continuing oversight. Applied to a department store, these principles favour a bounded and verifiable first project rather than a broad autonomous system.17

An associate-facing product-and-policy assistant fits those conditions. Three things already on the record point to it: the NRF department-store panel named product assortments, databases, data quality and taxonomy as the sector's most automatable complexity, the same panel cited a 90% reduction in the cost of product data management, and the quoted IADS member has shown that a department store can take ownership of that data layer without outsourcing it. It can answer questions about products, services, returns, delivery, appointments and approved procedures. Its answers can be grounded in approved product, service and policy information, the source can be shown to the associate, and unresolved questions can be escalated.

Lowe’s provides an adjacent retail example. Its Mylow Companion was deployed across more than 1,700 stores to give associates access to product, inventory and project information. A department-store version would have to accommodate the additional complexity of concessions, services, appointments, returns and store-specific policies.18

Preparing that knowledge is where the pilot becomes difficult. The IADS member treated data quality as a question of commercial ownership rather than a purely technical clean-up exercise, moving responsibility from IT to its e-commerce specialists and retaining human validation. The same issue would arise in an associate assistant: POS, CRM, inventory and product systems may disagree over whether an item is available, which department or concession owns the sale, or how a return should be classified. Store policies can also differ by service, market and location. The department store therefore has to decide which definition takes precedence, who owns it and how quickly an outdated answer is removed. None of this is free. Reconciling definitions, keeping formative work in human hands and staffing real oversight are costs the store carries before any savings appears, and they compete for the same budget as stores, people and logistics.

NIST requires systems to be tested before and during operation and reassessed against their intended purpose. For a first department-store pilot, that means testing the same set of real associate questions across the viable local, cloud, hybrid and non-AI options. The comparison should measure answer quality, response time, cost per accepted answer, escalation and manual correction, while also recording employee rework and variation across stores and languages.

Measure rework, not activity

The pilot should also test whether experienced employees consider the answers commercially useful, whether newer employees can explain why an answer is correct and whether the assistant builds understanding rather than conceals gaps.

Oversight should follow the retail decision itself, applying the NIST requirements set out above. An associate needs to see the product or policy source behind an answer and know when to escalate it. A buyer reviewing an AI recommendation needs the sales, returns and stock evidence behind it. A store manager needs the authority to reject a staffing recommendation that meets a labour target but leaves the floor unable to serve customers. Decisions affecting recruitment, promotion or discipline require stronger human control again.

Scale only the question types the pilot answered reliably, and leave the rest with people. The test of the pilot is whether associates keep using it once the launch attention fades, whether the cost per accepted answer is one the business would pay at scale, and whether the people supervising it can explain why an answer was right.

Owning the intelligence layer buys one thing above all: the freedom to replace the model.

A department store that keeps its commercial definitions, evaluation standards, workflows and customer identity separate from a supplier can move as the technology moves. It can adopt a better model, bring a sensitive workload closer to home or renegotiate with a provider without reconstructing the business around that decision. Once those assets are embedded inside one platform, switching becomes slower, more expensive and more dependent on the supplier.

Homegrown AI should preserve that freedom. Where a workload runs can change with economics, sensitivity and control. The product history, customer logic, evaluation standards and operating definitions that give the business its meaning must remain portable and under the department store’s control. The model can change. The department store should not have to rebuild itself around every change.

Credits: IADS (Maya Sankoh)

What: Airlines are adopting retail-style loyalty mechanics — gamified rewards, no-purchase engagement, and partner-driven monetisation — to keep customers connected between trips.

Why it is important: Airlines are proving that gamification and partner-driven rewards can turn loyalty into a revenue line in its own right, reinforcing the case for department stores to treat loyalty as a business rather than a cost centre.

Airlines are increasingly borrowing from retail's loyalty playbook, creating reasons for customers to engage with their brands outside active travel. TUI's new Smiles Rewards Club guarantees a monthly win without requiring a booking, while British Airways' Avios currency now draws more than 80% of its revenue from external partners such as Amex and JPMorgan Chase, rather than from flying itself. Virgin Atlantic frames the shift as building lasting relationships through experiences loyal members could not otherwise buy.

Personalisation and reduced friction are central to this shift. British Airways has replaced boarding passes with Face ID on some routes and now allows passport scanning instead of manual checks, while both airlines stress that data should surface relevant offers rather than drive engagement for its own sake. Jet2's myJet2Perks and Riyadh Air's Sfeer scheme extend the same logic without a formal points currency, offering discounts and non-expiring rewards regardless of whether a trip is booked.

The push is partly generational: OAG research shows only 65% of Gen Z travellers are enrolled in an airline loyalty scheme, against 89% of Baby Boomers, yet Gen Z redeems rewards more often while staying loyal to fewer providers, booking across multiple airlines on price or route.

IADS Notes:  Loyalty programmes across department stores and multi-brand retailers have been moving in a similar direction over the past several months. Reporting from Forbes in February 2026 found that nearly a third of Sephora's 46 million loyalty members now engage through gamified shopping mechanics, a precedent for the guaranteed-win, no-purchase-required model TUI has since applied in travel. Fashion Network noted in April 2026 that M&S had overhauled its Sparks scheme around spendable cash rewards, AI-driven personalisation and partner-based bonuses, including one tied to Virgin Atlantic bookings, closely mirroring the airline-retail crossover now emerging in loyalty economics. Fashion United reported in January 2026 that Harvey Nichols had restructured its own loyalty tiers as part of a wider brand repositioning, tying rewards innovation directly to broader transformation efforts. And Retail Dive covered, in June 2026, the linking of J.C. Penney's and Aéropostale's loyalty programmes under the Catalyst Brands portfolio, a coalition-style structure comparable to the cross-brand points transfers airlines run with retail and financial partners.

Miles better: the new business of airline loyalty

What: Southeast Asia split into two speeds in Q2 2026, with tech- and trade-led economies such as Vietnam, Malaysia and Singapore accelerating while Thailand and the Philippines lost momentum amid a Middle East-driven energy shock.

Why it is important: Vietnam and Singapore's continued strength — despite regional headwinds — reinforces their standing as priority growth markets for retail investment and expansion.

Southeast Asia's six major economies diverged sharply in the second quarter of 2026. Vietnam led with 8.39% GDP growth, followed by Malaysia at 6.0%, Singapore at 5.9% and Indonesia at 5.29%, while Thailand slowed to 1.9% and the Philippines to 2.3%. Exports were the standout driver across all six markets, powered by global demand for electronics and technology-related goods, with industrial activity strengthening in Malaysia, Singapore, Vietnam and the Philippines.

The Middle East energy shock complicated the picture, pushing inflation higher in five of the six economies and pressuring the Indonesian rupiah, Philippine peso and Thai baht. Policy responses diverged accordingly: Indonesia and the Philippines raised interest rates, Singapore tightened its exchange-rate policy, while Malaysia, Thailand and Vietnam held their settings steady.

Looking ahead, technology-led exports and improving manufacturing indicators provide a strong platform for the second half, but how effectively individual economies translate that momentum into broader domestic growth will depend on how inflation and currency pressures evolve.

IADS Notes: The World Bank's downgraded 2026 growth outlook for Southeast Asia had already begun weighing on major retail conglomerates by early spring, with stagnant sales and softer tourist arrivals pressuring groups such as Central Retail, Makro-Lotus and Big C (The Diplomat, April 2026). Against this backdrop, the region's AI adoption has produced uneven results: early movers have captured measurable efficiency and service-quality gains, but regulatory fragmentation and workforce readiness continue to limit how many retailers can scale AI beyond pilot projects (The Diplomat, April 2026). Performance has nonetheless diverged sharply by market. Singapore's retail sector kept accelerating even as petrol and living costs climbed, reinforcing its position as a regional benchmark for resilience (Inside Retail, June 2026), while Vietnam's retail sales moved into double digits on the back of a tourism boom and strong GDP growth, prompting Central Retail to commit to 30 new stores in the market (Inside Retail, April 2026).

Southeast Asia quarterly economic review: Tech tailwinds drive a two-speed region