A new global study of 240 senior procurement leaders, conducted by Foundry (IDG) for CIO Market Pulse, measures the gap between agentic AI ambition and execution, and names what’s actually closing it.
TL;DR
- 82% of procurement leaders are open to AI agents negotiating purchases on their behalf. Only 6% have built it.
- The vast majority of organizations have automation. Almost none have autonomy. The difference is doing more damage than most CPOs realize.
- The two leading barriers to scaling AI in procurement are not technology problems. They are governance ones, and they rank ahead of the items dominating most weekly status reports.
- A leading cohort of procurement organizations is committing to orchestrated agentic workflows or autonomous operations in the next 24 months. The rest will spend that window catching up.
- The report flags its own headline maturity number as an aspirational upper bound, not a verified count. It is the kind of self-disclosure that vendor-funded research rarely makes.
- A new global benchmark of 240 senior procurement leaders, conducted by Foundry (IDG), is now available. Read the full report.
The Paradox at the Center of Procurement in 2026
Procurement’s AI problem in 2026 is not appetite. It is delivery.
A new global study of 240 senior procurement leaders (CPOs, VPs, Directors, and Heads of Procurement at enterprises across Europe, North America, and APAC) measures the gap precisely. Eighty-two percent are open to AI agents negotiating purchases on their behalf. Six percent are already doing it. A small minority have reached governed end-to-end autonomous operations.
The demand signal is overwhelming. The execution signal is barely audible. And the barriers between the two are governance ones, not technology ones. They rank ahead of ERP integration, data quality, skills gaps, and ROI uncertainty, the items dominating most weekly status reports.
The pattern fits a wider one. Gartner expects more than 40% of agentic AI projects to be canceled by the end of 2027, citing the same execution gap that procurement is now living. The Foundry data shows what that gap looks like inside procurement specifically.
Why This Research Exists, and Why It Matters Now
The agentic procurement conversation in 2026 has been carried, until now, by vendor claims and analyst forecasts. Both have their value. Neither is the same as procurement leaders speaking for themselves.
This is one of the first large-scale buyer-side benchmarks of agentic procurement to be published in 2026. Conducted by Foundry (IDG) between January and February 2026, the study qualified respondents on enterprise revenue thresholds (USD 500M+ in APAC and EMEA, USD 750M+ in North America) to ensure that only enterprise-scale procurement environments were measured. Sixty-three percent are primary decision-makers. Fifty-nine percent are CPO-to-Director level. Testing runs at 90% confidence intervals.
What that means in practice: the methodology stands up to board-level scrutiny. That is a bar most agentic procurement coverage in 2026 has not attempted to clear.
The Gap the Report Names
There is a difference between automation and autonomy. Most organizations have the first. Almost none have the second.
The two words sound similar. The architecture behind them is not. Automation makes individual procurement steps faster and cheaper. Autonomy strings those steps together into a workflow that runs itself inside set boundaries, with humans drawn in only when the situation requires judgment. The report measures where the market sits on that spectrum across seven core procurement functions, from intake and request routing through to invoice processing. The pattern is consistent. The modal position is not autonomy. It is one stage short of it, and the report places the bulk of the market there with a single figure.
What stands out is the report’s intellectual honesty about its own headline number. The figure for organizations claiming end-to-end autonomy is explicitly flagged inside the research as an aspirational upper bound, not a verified count. The supporting data on governance readiness and outcome attribution does not match it. Vendor-funded research rarely discounts its own most quotable stat. This one does, and that single editorial decision tells you what kind of report this is before you read a single chart.
For every CPO running an AI program right now: the report gives you a four-stage maturity diagnostic that lets you place your organization honestly on the curve in under five minutes. That diagnostic alone makes the research worth reading.
What Closes the Gap Is Not What Most Organizations Are Trying
Asked how they plan to address security and trust concerns, most procurement leaders name a technology project. Better integrations. Stronger model validation. Tighter access controls. More data hygiene. The report measures all of them and finds that none is the limiting factor.
What separates organizations that scale AI from those that stall is governance built into the execution layer itself, not as policy, not as oversight, but as the architecture inside which AI runs. McKinsey’s State of AI research has consistently identified governance, risk, and trust as the leading constraints on scaling generative AI in the enterprise, ahead of technology or ROI concerns. The Foundry data quantifies what that constraint looks like in procurement.
The report puts one figure at the center of the discussion: a single trust metric, specific and uncomfortably low, that determines whether organizations can act on AI outputs at scale or whether every decision still gets manually re-verified. That metric, more than any other in the study, separates real autonomy from a more expensive version of the manual process AI was meant to replace.
“We have shifted from manual tail-spend firefighting to strategic, large-scale negotiations, speeding up decisions, maximizing savings, and enabling our procurement team to focus on higher-priority tasks.”
— Jaswinder Saini, VP and Head of Supply Chain, as cited in the Foundry/CIO Market Pulse research
What the Report Will Answer for You
The full report walks through seven sections of evidence. Four questions sit at the center of it:
- Where the market stands on agentic procurement maturity, and how to honestly place your own organization on the curve
- Which procurement function is the most-wanted, least-built use case for agentic AI in 2026
- What separates the cohort committing to orchestrated autonomy from those still experimenting
- The single architectural shift that closes the ambition-execution gap, drawn directly from the data
These are the four questions the report answers, with the specific numbers, cohorts, and architectural shift behind each.
Why the Next Twenty-Four Months Will Sort the Market
Pullback from AI is rare in the data. Only a small fraction of respondents plan to reduce AI investment. But the rest of the market is not moving at the same speed.
A leading cohort is committing, explicitly and with budget, to either orchestrated agentic workflows or fully autonomous procurement operations over the next 24 months. The majority are still running pilots, expanding copilots, or watching. The report quantifies the gap between the two groups, and it is larger than the consensus suggests.
Every month the leading cohort operates at higher autonomy is a month the rest of the market is not. By the time the trailing group catches up, the leading cohort will have moved again.
The procurement function that comes out of 2026 looking like the leader will not be the one that experimented the most. It will be the one that picked the right operating model early, instrumented it properly, and committed.
The report names what the leading cohort is doing differently, where the rest of the market is spending in the wrong places, and the specific moves that separate the two.
Read the Full research
Most Procurement AI Investments Are Stalling. Here Is Why, and What to Do.
A global study of 240 senior procurement leaders. Research conducted by Foundry (IDG) for CIO Market Pulse. Sponsored by Zycus.
It is the kind of self-disclosure that vendor-funded research rarely makes.
Related Reads:
- From Co-Pilots to Commanders: How Agentic AI is Redefining Procurement Transformation
- Why Agentic AI Is the Future of Source-to-Pay Automation by 2026
- The Complete Guide to Agentic AI in Procurement
- Whitepaper: Beyond Integration – 8 Reasons to Choose an AI-Agent Orchestrated S2P Suite
- From ‘Cognitive Procurement’ to Agentic AI — What Actually Changed
