Why more data and AI tools aren't the answer

This blog is part of a series from Closing the Gap: How Procurement Moves from Insight to Action, a whitepaper produced by Beroe in partnership with Kearney. The series explores why procurement’s operating model is struggling to keep pace with its expanded mandate, and what it will take to close the gap between intelligence and action. 

In our previous post, we looked at why AI has changed what procurement produces without changing how procurement decides. That raises an obvious question: if AI alone hasn’t solved the problem, what will? For a lot of organizations, the instinct is to keep adding, another dashboard, another data source, another specialized tool. This post looks at why that instinct usually makes things more complicated instead of simpler.

Progress that doesn’t show up in decisions

Procurement is undeniably more digital than it was a decade ago. Spend visibility has improved, supplier risk monitoring is now standard practice, market intelligence is far more detailed, and analytics teams have grown as AI has worked its way into daily workflows. Execution has been digitized too, through source-to-pay platforms, supplier management tools, and workflow automation. 

On the surface, that looks like transformation. What has actually emerged is two capable pillars, intelligence on one side, execution on the other, built up independently of each other. Between insight and decision is a gap, and the significant work that needs to happen to address this gap remains inconsistent and largely manual.

The hidden factory problem

One of procurement’s most persistent constraints is what’s sometimes called the “hidden factory,” all the informal, manual work that sits outside formal processes but is quietly needed to keep those processes running. In procurement, hidden factories are everywhere. 

They show up in offline spreadsheets used to reconcile spend data, in supplier trackers that only one person really knows how to maintain, in email threads where exceptions get agreed but never make it into any formal system, and in slide decks that hold the real category strategy while the official system only stores a template. 

These workarounds exist for a good reason. Procurement professionals build them because formal systems don’t fully capture the complexity of their work. The trouble is that once a workaround becomes embedded, it becomes part of how the function runs in reality. Over time, that makes procurement harder to scale, because the knowledge that keeps things moving sits with individuals rather than with any system that could pass it on.

The threshold problem

A related constraint is the threshold for action. Most procurement organizations run on thresholds, whether they are written down anywhere or not. A sourcing event has to be large enough to justify the effort. A savings opportunity has to be material enough to make it into the pipeline. A risk has to be severe enough to escalate. A contract variance has to be large enough to review. A category has to be strategic enough to earn senior attention. 

Thresholds exist for a sensible reason: human capacity is finite. But they also leave value on the table. In a manual operating model, procurement naturally gravitates toward the big opportunities, because no team can chase every small negotiation, every minor cost movement, every emerging supplier risk, or every localized pricing advantage. The cost of investigating and acting on something small often outweighs the value it would deliver. 

This is one area where AI and continuous decisioning genuinely change the economics. If a system can identify, qualify, prepare, and in some cases even execute lower-value opportunities at very low marginal cost, the threshold for action can drop. A single $50,000 opportunity may never justify a category manager’s time on its own. Hundreds of them across a global enterprise can add up to real margin, provided there is a system built to find and act on them.

The AI productivity illusion 

When teams start using generative AI, the productivity gains show up fast. Reports come together quicker, summaries take minutes instead of days, and everyone can point to time saved. The risk is mistaking that speed for better decision-making. A category strategy, a supplier risk summary, or a market report drafted faster is not automatically a better one. 

AI can also produce more content than a team can take in. When every user can generate reports, summaries, scenarios, and recommendations on demand, organizations can end up facing a new kind of information overload, just a faster-moving version of the old one. More AI-generated content simply becomes more noise unless it is tied to clear decision logic, a level of confidence, accountability, and an actual action to take.

Measuring the wrong thing

Part of why this illusion persists comes down to how procurement transformation gets measured. Platform adoption, data coverage, and supplier monitoring are the metrics that tend to show up on a transformation scorecard, and they are useful as far as they go. None of them prove that decision quality has improved. 

A more outcome-led approach asks different questions. How quickly did the team move from signal to action? How much value did it capture that would previously have slipped through? How much risk got mitigated before it turned into disruption? How often did an insight actually lead to something being done about it? 

This matters because what procurement measures shapes how procurement behaves. A function built to track process completion will get very good at completing processes. A function built to track realized value, decision speed, risk mitigation, and market outperformance will get better at the thing that actually counts: making good decisions, often enough and fast enough for them to matter. 

More data and more tools were never going to close this gap on their own; procurement needs a way to connect what it already has, its data, its intelligence, and its execution layers, into something that consistently produces decisions.  

This article focused on the problems – the solution is what we turn to in the next part of this blog series drawn from Closing the Gap: How Procurement Moves from Insight to Action, a whitepaper authored by Beroe and Kearney.

If you’re looking to find out more about the missing layer between insight and action now, you can find it in the whitepaper.