For two decades, procurement technology sold itself on one promise: automate the transaction. Replace the paper PO with a digital one. Replace the manual three-way match with a system-driven one.

‍

That promise mattered, and most enterprises are still only partway through cashing it in. But HFS Research’s 2026 analysis of the source-to-pay market makes a sharper claim — automation of the transaction is no longer where the competitive advantage lives. It’s table stakes. The real disruption is happening one layer up, in how spend decisions get made in the first place.

‍

‍

The Claim, Precisely

HFS Research’s Hridika Biswas frames the shift directly: S2P is no longer just about digitizing procurement — it’s evolving into a question of who controls how spend decisions get made. The market is moving away from downstream, transaction-heavy automation toward upstream, intelligence-led orchestration, where category strategy, supplier risk, and demand shaping are what actually drive value.

‍

What used to differentiate vendors — automation, P2P efficiency, platform rollout — is now simply the baseline. The disruption is happening in the layers that were historically under-digitized: sourcing, supplier intelligence, and decision-making itself.

‍

That’s a meaningfully different way of describing the procurement technology market than most vendors use, and it’s worth sitting with because it reframes what “good” procurement technology should actually be evaluated on.

‍

‍

Why This Matters More Than It Sounds Like It Does

If automation of P2P transactions is table stakes, then a platform that only does that — processes invoices, routes approvals, executes payments — is competing on a dimension that no longer differentiates. Every serious S2P vendor can do this now. The question a CPO should actually be asking a vendor isn’t “can you automate my invoice processing,” but “can you give me better sourcing decisions, earlier supplier risk signals, and demand visibility I don’t currently have.”

‍

HFS Research’s broader 2026 survey work backs this up with hard numbers: enterprises that have moved toward orchestrated, AI-enabled procurement are seeing measurable impact — 3% to 10% in cost savings, 90% compliance, increased spend under management, and real-time spend visibility. Notice what’s on that list.

‍

Compliance and visibility aren’t “nice to have” governance metrics anymore — they’re being delivered as a direct output of the same orchestration layer that drives savings. That’s the upstream/downstream distinction in practice: the orchestration layer doesn’t just process transactions faster, it changes what procurement can see and decide before the transaction even happens.

‍

‍

Where the Old Model Breaks Down

The transactional model assumes procurement’s job is to execute a decision someone else already made — a requisition gets approved, procurement sources it, the order gets placed. The intelligence-led model assumes procurement’s job is to shape the decision itself — flagging that a category is trending toward supply risk before a stockout happens, or surfacing that three business units are sourcing the same indirect category from different suppliers at different price points.

‍

This is also where most enterprises are stuck. HFS Research is explicit that this transition is not yet complete: enterprises are not fully ready to realize the potential of AI-driven orchestration because of a lack of structured data, siloed platforms and workflows, and limited knowledge sharing among stakeholders. In other words, the technology to enable intelligence-led procurement exists. The organizational and data readiness to actually use it often doesn’t — yet.

‍

‍

What This Means for How You Should Evaluate a Platform

If HFS Research’s framing is right — and the direction of the market suggests it is — then the evaluation criteria for an S2P platform should shift accordingly:

‍

Does it unify data, or just digitize a form? A platform that automates onboarding but still leaves supplier data fragmented across three systems hasn’t actually solved the upstream problem. Unified, ERP-agnostic supplier data is the precondition for any intelligence layer working at all.

‍

Does it surface decisions, or just execute them? Spend analytics that tell you what already happened is useful for reporting. Spend analytics with predictive intelligence — flagging a category trending toward risk, or a supplier whose performance is degrading — is what actually changes a decision before it’s made.

‍

Does compliance happen automatically, or does someone have to chase it? HFS Research’s 90% compliance figure isn’t incidental to the orchestration story — it’s a direct consequence of moving compliance checks into the workflow itself rather than treating them as a separate audit function.

‍

The takeaway for procurement leaders isn’t that transactional automation was wrong to invest in — it’s the necessary foundation. But foundation isn’t where the next round of value gets created. That’s happening in the layer above it, exactly where HFS Research says to look.

‍

‍

‍

Source: HFS Research, “CPOs Must Unite Enterprise Processes, Platforms, and People for AI-Enabled Outcomes,” 2026.

‍