What: BCG’s procurement study shows that agentic AI delivers stronger results when companies combine deployment with governance, capability building and process redesign.

Why it is important: This matters because agentic AI can improve procurement performance only when organisations redesign workflows, strengthen data foundations and build trust in autonomous decision-making.

BCG’s 2026 tech procurement study finds that most enterprises are already piloting or deploying agentic AI, shifting the question from whether to use the technology to how procurement organisations should adapt around it. Based on a survey of more than 200 CIOs, procurement leaders and technology buyers, the research shows that internal operational benefits usually appear before supplier-facing commercial gains. Agentic AI is already improving speed, throughput, manual effort and process discipline, but stronger negotiation outcomes, reduced vendor lock-in and better supplier performance require broader organisational change. Production deployments outperform pilots because value emerges when companies connect workflows and redesign the operating model, rather than testing isolated use cases. BCG identifies four main barriers to scale: inconsistent data, legacy-system integration, unchanged processes and governance constraints. Trust in autonomous decision-making is the biggest organisational barrier, cited by 71% of respondents, followed by security and IP risks at 66%. Procurement leaders should redesign processes first, build capabilities alongside technology, measure both operational and commercial outcomes, strengthen data foundations and redefine procurement around market intelligence and commercial judgment.

IADS Notes: Recent industry analysis shows that agentic AI in procurement is moving from experimentation toward organisational redesign. In June 2026, Journal du Net warned that procurement AI becomes risky without governance, high-quality data, human oversight and clear accountability. BCG’s September 2025 work on supplier negotiations showed that GenAI can improve cost analysis, risk reduction and negotiation support, but only when integrated across end-to-end operations. McKinsey’s January 2026 analysis of AI agents and ERP systems reinforced that value depends on connecting agents to core business processes rather than leaving them isolated from operational systems. BCG’s July 2026 work on decision agents similarly stressed the need for governance, cross-functional data layers and structural investment, while BCG’s June 2026 research on agentic AI as a transformation engine showed that teams capture value when they redesign how they operate around the technology.

Scaling agentic AI in procurement is an organisational challenge