What: A survey of The Information's subscribers shows Claude and Gemini gaining on ChatGPT over the past year, with Claude Code leading coding agents and 46% of respondents using some open-source models.

Why it is important: Leadership among AI assistants and coding agents changed within a year, so a vendor choice made now may not hold, which is a concern for retailers that already trail other sectors on maturity, with retail scoring 49 out of 100 in BCG's index.

The Information surveyed its subscribers last month on enterprise AI. Nearly two-thirds said their organizations develop AI applications, and a third said they receive a return that is a multiple of what they spend. Another 17% report positive but lower-than-hoped returns, 30% say it is too soon to tell and 10% call the spending a waste. Last year's Massachusetts Institute of Technology study found only 5% of custom enterprise AI tools reached production.

On tools, usage of Claude and Gemini rose over the year at the expense of ChatGPT, which still leads and has regained some ground. Microsoft Copilot usage is falling. Among coding assistants, Claude Code leads, OpenAI's Codex has surged and Cursor is not far behind. About a quarter use Perplexity or DeepSeek, and a sixth use Moonshot's Kimi K3. Some 46% use open-source models and 21% Chinese ones, yet only 20% run most workloads on open source.

A year ago, 59% of readers had never used an AI agent. On hiring, 22% are hiring fewer workers against 10% hiring more, while nearly half report no change.

IADS Notes: The survey's finding that a third of respondents see returns that are a multiple of their AI spend contrasts with the more cautious picture in earlier coverage. BCG's Applied AI Index found almost half of companies generating measurable value, concentrated in a 7.5% group of "future-built" companies (BCG, September 2026), and its CEO survey showed that nearly nine in ten CEOs report targeted gains while only 14% have defined the financial impact of every AI project (BCG, July 2026). In retail, AI spending is rising but adoption remains broad and shallow, with Levi Strauss running around 1,000 agents across supply chain, planning and wholesale, an early example of the back-office agent deployment the survey describes (WWD, July 2026). The survey's rise of Chinese open-weight models ties into BCG's analysis of US and Chinese AI stacks separating, which turns model and vendor choices into a regulatory and geopolitical question (BCG, June 2026). The shifting assistant landscape matters for brands, which now appear inside ChatGPT and Claude on the platforms' terms, controlling their content but not the ranking or the conversation (Journal du Net, September 2026). On staffing, the surveyed subscribers mostly report unchanged hiring, whereas companies in BCG's index expect AI to cut headcount by roughly 10% to 15%, concentrated in middle management (BCG, September 2026).

Subscribers to the tech review The Information report AI returns rising as Claude and Gemini gain on ChatGPT