Originally aired December 4, 2025 on Extropy On AIr.
AI in procurement isn’t coming, it is already here. The question is how to move past basic automation to real impact. In the final episode of our Raindrop partner series, Kevin Zirkle of Extropy Advisors and Ward Karson, COO of Raindrop, explore the three stages of AI maturity, where AI is already delivering value across sourcing and spend management, how autonomous sourcing will work, and how CFOs can evaluate and trust AI.
Listen: Extropy On AIr podcast page | Spotify
Key takeaways
- Most companies are still in basic automation. True AI adoption means predictive insight and decision support.
- AI maturity moves through three stages: automation, augmentation and autonomy. Most teams today are in augmentation.
- AI is already at work across contracts, supplier risk, sourcing, e-procurement and analytics, including plain-English questions about spend.
- Autonomous sourcing, starting with tail spend, will let procurement teams focus on strategy and supplier relationships.
- Evaluate AI on business value first, governance second and the right implementation partner third.
Where are most companies with AI in procurement?
Ward says most companies are still focused on basic automation: moving paper and email processes into digital forms. That is a good start, but it isn’t intelligence. True AI adoption means predictive insights, decision support and dynamic optimization. A few leaders are making that leap. And despite the hype about AI agents doing everything, he says they won’t take most jobs. They will take away the mundane, rote work.
Kevin frames it as three stages of AI maturity:
- Automation: digitizing and simplifying steps.
- Augmentation: helping people make better decisions.
- Autonomy: letting systems make decisions within safe boundaries.
Ward agrees. Most Raindrop clients are solidly in augmentation, moving past manual workarounds toward systems that flag risks, recommend actions and surface patterns. Kevin calls this the stage where the magic happens, because people stay in control while their speed and accuracy improve dramatically.
Where is AI already delivering value in procurement?
Ward walked through the Raindrop platform:
- Contract management: extracting terms, suggesting clauses and flagging risky language.
- Supplier management: dynamically scoring supplier risk, ESG exposure and financial health.
- Sourcing: recommending suppliers, auto-scoring bids, flagging anomalies and guiding suppliers on pricing to drive cost savings.
- E-procurement: guiding purchases, catching duplicates and enforcing policy.
- Analytics: predictive forecasts and natural language questions.
That last one is a big deal for CFOs and controllers. With no training, users can ask questions like “Show me suppliers with increasing spend in Q2,” “Which sourcing events saved more than $100,000 last year?” or “What contracts expire in the next 90 days with suppliers rated high risk?” Kevin notes it reduces dependence on data analysts and makes procurement intelligence accessible to executives.
Behind it, Raindrop builds its AI on a graph-based data model that captures the relationships between suppliers, contracts and events, then layers on use-case-specific models, from large language models for natural language to predictive models for sourcing outcomes and supplier risk. Everything is explainable, so users can see what drove a suggestion and override it. Kevin points out that CFOs need to audit not just the outcome but how the system got there.
What will AI in procurement look like in three to five years?
Ward sees three big shifts:
- Autonomous sourcing. AI will run RFPs end to end, selecting suppliers, issuing events, negotiating terms and awarding contracts based on rules and past outcomes, with people reviewing only exceptions.
- Proactive spend optimization. AI will spot inefficiencies in real time and suggest consolidation, renegotiation or supplier changes before the spend happens, not after.
- Embedded and custom agents. Platforms will offer built-in agents based on best practices, and companies with in-house AI skills will be able to build their own on the same platform.
Tail spend is where autonomous sourcing is starting. These high-volume, low-dollar events rarely get attention because sourcing teams focus on their top suppliers. Raindrop can auto-initiate the event, select pre-vetted suppliers, generate the RFP, score responses and recommend an award without human involvement, unless the risk score crosses a set threshold. That frees teams for innovation, supplier relationships and long-term value.
Raindrop is also investing in risk signal ingestion, pulling in third-party data like ratings, ownership registries and emissions sources so AI can continuously score suppliers and flag concerns. Kevin sees that kind of proactive ESG monitoring becoming something investors expect.
What results can AI-driven sourcing deliver?
Ward shared the example of a global technology company that has automated more than 80% of its sourcing events with Raindrop’s AI. The system scores bids, recommends suppliers, enforces compliance and even sends award notices. Procurement now spends its time on supplier innovation and strategy instead of tactical execution. As Kevin puts it, this isn’t ten years out. It is happening now.
How should CFOs evaluate AI in spend management?
- Business value first. Don’t chase AI to keep up with the Joneses. Start where decisions are being made manually or slowly, and ask whether intelligence could help.
- Governance. Make sure you have controls for data privacy, model transparency and auditability. Ward recommends starting with a responsible AI charter that defines where AI helps, who owns what and how you will monitor performance.
- Partnerships. Work with platforms and firms that understand not just the technology, but how to implement and use it in the real world.
Kevin adds that an AI investment isn’t a technology decision, it is an operating model decision. You are deciding how your people and processes will work five years from now.
How do you build trust in AI outputs?
Ward points to two things. First, transparency: every AI output in Raindrop comes with a confidence score and a trail showing what data and rules were used. Second, validation: clients can test AI outcomes against human decisions to see where they align and where AI improves results. Kevin calls that kind of instrumentation essential if AI is going to earn a seat at the CFO’s table.
In their words
Business value has to be first, always.
Ward Karson, Raindrop
An AI investment isn’t a technology decision. It’s an operating model decision.
Kevin Zirkle, Extropy Advisors
About the speakers
Kevin Zirkle is Director of Business Development at Extropy Advisors, where he works across Extropy’s partner ecosystem, matching organizations with the right technology and the right people to move transformation from the whiteboard into production.
Ward Karson is Chief Operating Officer at Raindrop, a spend management platform covering sourcing, contracts, supplier management, e-procurement and analytics.
FAQ
What are the stages of AI maturity in procurement?
There are three: automation, which digitizes and simplifies steps; augmentation, which helps people make better decisions by flagging risks, recommending actions and surfacing patterns; and autonomy, where systems make decisions within safe boundaries. Most procurement teams today are in the augmentation stage.
What is autonomous sourcing?
Autonomous sourcing is when AI runs sourcing events end to end: selecting suppliers, issuing RFPs, scoring responses and recommending or making awards based on rules and past outcomes, with people stepping in only for exceptions. Tail spend, with its high volume of low-dollar events, is one of the best places to start.
How can CFOs trust AI-driven procurement decisions?
Look for transparency and validation. Every AI output should come with a confidence score and a trail showing what data and rules were used, and teams should be able to test AI outcomes against human decisions. Starting with a responsible AI charter that defines where AI helps, who owns what and how performance is monitored also builds trust.
Keep exploring
- More from our partner series on Extropy On AIr
- Previous in the Raindrop series: Compliance, Visibility, and Control with Raindrop
- Related: AI, Insight & the Future of Finance with Esker
- Why Extropy
Ready to move from AI pilots to real procurement impact? Book a discovery call with our team.
