AI in procurement has moved from assistants that answer questions to agents that take steps in the process. This guide explains what changed, what still holds teams back, and how to test what vendors claim.
If you are ready to compare vendors, go to our list of the best AI procurement platforms.
Key Takeaways
Most procurement leaders now expect AI to change their function. The Hackett Group reports: “Eighty percent of procurement executives now identify AI-enabled technology as the most transformational trend affecting the function over the next five years” (Hackett, March 17, 2026). It also notes that current AI use “is concentrated in areas such as contract management, market intelligence, and spend analytics.”
In 2026, a platform is AI-powered when AI does real work in the process, not only in a dashboard. Look for these signs:
The first wave of AI in procurement was assistants: chat tools that answer questions or draft text when a person asks. The current wave is agents: AI that takes steps in a workflow, such as collecting missing details, routing a request or updating a record, and hands decisions with risk to a person.
For the difference in detail, see what makes an AI procurement platform agentic. For how agents work from request to payment, see agentic AI in procure-to-pay.
The payoff can be large for teams that get it right. Deloitte found that “Digital Masters” put up to 24% of their budgets into technology, and report an average 3.2x return on GenAI, compared with a projected return slightly above 1.5x for followers (Deloitte, August 2025).
Adoption is wide but shallow. Hackett reports that 43% of organizations are actively pursuing AI deployment, but only 12% report large-scale implementation (Hackett, March 2026).
People and process are part of the gap. Gartner reports: “Sixty-three percent of senior procurement leaders expect AI to significantly improve procurement performance over the next three years,” yet only 25% have undertaken job and role redesign (Gartner, October 8, 2026).
For reference, here is what Opstream shares about its own results: customers report 73% faster request handling, and Opstream scores 9.7/10 ROI (G2 / self-reported).
To compare vendors, see our list of the best AI procurement platforms, or see how Opstream applies AI for procurement.
What is agentic AI in procurement?
Agentic AI in procurement means AI agents that can take steps in a process, not only answer questions. An agent can collect missing details from a requester, check a request against existing contracts, prepare a summary for reviewers and move routine work forward, while people approve the decisions that carry risk. See what makes an AI procurement platform agentic.
How is an AI procurement platform different from an S2P suite with AI added?
An AI procurement platform is built around AI from the first request: it uses AI to capture needs, route work and act across the systems you already run. An S2P suite with AI added usually puts AI features inside its own modules, such as invoicing or sourcing. Ask each vendor which steps the AI takes on its own, which systems it can act in, and how a person reviews its work.
Why do most procurement AI projects stall at pilot?
The Hackett Group found that 43% of organizations are actively pursuing AI deployment, but only 12% report large-scale implementation (Hackett, March 2026). Common reasons are poor data, AI that sits outside the daily workflow, and roles that were never redesigned. Gartner found that only 25% of senior procurement leaders have undertaken job and role redesign (Gartner, October 2026).
Want to see AI agents work on your own requests? See how Opstream uses AI for procurement.
About the Author
Lihi Lutan is the Co-Founder and CEO of Opstream, changing the way companies buy. Throughout her career, Lihi built and scaled business operations at startups and large corporations. Early in her career, Lihi was with Cyota (acq. RSA Security) as a team leader and project manager before moving to Thomson Reuters and Fundtech to manage global projects. Later, Lihi joined Taboola (NSDQ: TBLA) as employee 15, as VP Professional Services and Operations, leading the department as the company scaled from $8M to $1B in revenue. Transitioning from Taboola to StokeTalent (acq. Fiverr), Lihi served as the company’s COO. Lihi holds an LLB of Law and BSc of Computer Science from Tel Aviv University.
Connect on LinkedIn →References
1. The Hackett Group, press release on procurement’s AI agenda, March 17, 2026. thehackettgroup.com
2. Gartner, press release on CPOs and AI role redesign, October 8, 2026. gartner.com
3. Deloitte, 2025 Chief Procurement Officer Survey, August 19, 2025. deloitte.com