Agentic AI has moved from procurement buzzword to boardroom mandate in 2026. Gartner forecasts agentic AI spend in supply chain management software will grow from under US$2bn in 2025 to US$53bn by 2030, with Deloitte citing that 92% of CPOs are now assessing or piloting the technology.
Bain & Company highlighted the potential through an example involving an unnamed global bank. The bank developed an agentic AI solution to support its procurement team, using a conversational interface to guide employees through purchasing requests, capture and improve data and reduce manual effort across the buying process. Bain estimates that, once the solution is deployed at full scale, it could generate savings of up to US$180m.
Despite these successes, adoption remains uneven. Most procurement functions are still moving from pilot to production.
That gap between ambition and execution is the subject of The Future of Agentic AI panel at Procurement LIVE: The London Summit (8-9 September 2026, QEII Centre, Westminster), featuring Chief Procurement Officers from Kantar and Revolut, alongside senior transformation leaders from Unilever and London Luton Airport.
What agentic AI means for procurement
Agentic AI differs from the generative AI most procurement teams adopted first. Where generative AI drafts and summarises, agentic AI executes. Deloitte describes agentic systems as having governed access to enterprise systems, defined decision rights and the ability to adapt dynamically within guardrails, rather than simply following a script.
In practice, agents handle spend analysis by ingesting data across disconnected systems, run structured negotiations with suppliers, automate contract and compliance management, and process requisitions and purchase orders end-to-end. The technology also supports supplier selection through autonomous review of pricing history and sustainability risk.
While the benefits are clear, Gartner, Bain and McKinsey all highlight the same concern: the risk of moving too fast without governance. Gartner predicts that by 2028, 90% of B2B buying will be intermediated by AI agents, with more than $15 trillion of spend flowing through agent-mediated exchanges. At the same time, Gartner forecasts that over 40% of agentic AI projects will be cancelled by the end of 2027 due to escalating costs, unclear business value or inadequate risk controls.
Bain's research on financial-services procurement finds that organisations with more advanced, AI-enabled operating models have achieved staff efficiency gains of more than 30%, in some cases up to 50%, alongside savings of 5-8% on addressable third-party spend. Bain's broader work on agentic AI in procurement emphasises that effective AI deployment can raise procurement productivity by 60% or more and deliver incremental savings of 3-7%.
McKinsey's 2026 State of AI Trust research adds a governance point of view: almost two-thirds of respondents cite security and risk concerns as the top barrier to scaling, well ahead of regulatory uncertainty or technical limitations. The Hackett Group's 2025 Key Issues Study finds that while 49% of procurement teams piloted AI in 2024, only 4% reached large-scale deployment. For procurement leaders looking to close that gap, AI Procurement Courses offer a practical starting point for building the skills needed to move from experimentation to operational reality.
Real deployments, not just pilots
Several organisations represented at Procurement LIVE's Agentic AI panel are already running agentic systems at scale. Unilever is building a multi-agent procurement system on Google's Gemini Enterprise Agent Platform, splitting buying decisions across specialised agents that review supplier data, pricing history, contract terms and sustainability risk, before a coordinator agent presents a single view to procurement teams.
Kantar, a Bain Capital portfolio company, has adopted Globality's AI agent, known internally as 'Glo', to handle the analytical heavy lifting of evaluating sourcing options, comparing proposals and mapping supply markets.
"Recognising and respecting the human element of procurement is more important now than ever because many of the human contributions to supply networks [...] will increasingly become the domain of technologies like automation, robotics, and artificial intelligence," said Stephen Day, Chief Procurement Officer at Kantar.
Revolut has undertaken a procurement technology transformation with Coupa, consolidating spend visibility and controls across more than 20 markets as the fintech scales globally. "This transformation positions us to scale efficiently while maintaining the innovation and agility that defines Revolut," said Lauren Richards, Chief Procurement Officer at Revolut.
The Future of Agentic AI panel
The Future of Agentic AI panel takes place on Day One of Procurement LIVE, 15:50-16:30 (BST), in association with Zip. It will explore applications in spend analysis, supplier selection and process automation. Speakers include Nico Bac, Founder of Digital Procurement Now (Moderator); Philip Halanen, Head of Procurement at London Luton Airport; Alexey Gorchakov, Global Lead of Procurement Transformation and Technology and AI at Unilever; Stephen Day, Chief Procurement Officer at Kantar; and Lauren Richards, Chief Procurement Officer at Revolut.
Bain's research on financial-services procurement finds that organisations with more advanced, AI-enabled operating models have achieved staff efficiency gains of more than 30%, in some cases up to 50%, alongside savings of 5-8% on addressable third-party spend. AI for Executives & Strategy resources can help leadership teams assess where agentic AI fits within their own operating model before committing to large-scale deployments.
Why this matters for executives and strategy
The gap between pilot and production is the defining challenge for procurement leaders in 2026. Gartner's prediction that over 40% of agentic AI projects will be cancelled by the end of 2027 is a warning that deployment without governance is expensive. The organisations that reach scale - like Unilever's multi-agent system and Revolut's Coupa transformation - share a common approach: they redesign the function around AI rather than bolting it onto existing processes.
For executives, the takeaway is concrete. The data from Bain and McKinsey shows that the rewards for getting this right are measured in double-digit efficiency gains and 3-7% incremental savings. The costs of getting it wrong are cancelled projects and wasted spend. The question is not whether to adopt agentic AI, but how quickly your organisation can build the data quality, governance and talent to run it at scale.
Your membership also unlocks:
