Share with your COO
Procol is betting that procurement’s real inefficiency isn’t bad software, it’s too much software. At DPW Amsterdam 2026, the company will demo Clara, an AI orchestration layer that accepts plain-language requests and executes multi-step procurement workflows across SAP Ariba, Coupa, and Oracle without moving any data out of those systems. Early enterprise deployments, per Procol, show an 80 percent reduction in end-to-end procurement cycle times and a 15 percent drop in operational processing costs.
What this means for your business
The question this raises isn’t whether AI can automate procurement tasks. It’s whether your procurement function’s current bottleneck is execution speed or something harder, like supplier relationships, contract risk, or category expertise that no orchestration layer touches. If your team spends more than a third of its time on routing, approvals, and status-chasing across systems, Clara’s architecture is directly aimed at that waste. If the bottleneck is upstream judgment calls, this announcement is mostly noise for now.
The no-migration claim deserves scrutiny. Procol positions Clara as a layer on top of SAP, Coupa, Oracle, and existing ERPs, meaning it reads from and writes back to those systems without a rip-and-replace. That architecture is genuinely attractive to enterprises that spent years and significant capital implementing those platforms. The catch is that orchestrating across systems with different data models and permission structures is where these integrations typically degrade under real-world load. The 99 percent reduction in manual touchpoints is a striking figure, but it’s self-reported by a vendor, published by a construction-industry trade outlet with a natural interest in technology adoption stories, which means it reflects early deployments under favorable conditions and tells you nothing about the failure distribution.
The real strategic signal here is about where procurement sits in the enterprise hierarchy. Procurement has historically been underfunded on tooling relative to its spend influence. If agentic AI, where software agents autonomously execute multi-step tasks rather than just surfacing recommendations, actually compresses cycle times at the scale Procol claims, the COO case for upgrading procurement infrastructure gets considerably stronger. The leaders who win this bet early aren’t the ones who move fastest on Clara specifically. They’re the ones who’ve already standardized their ERP data well enough that an orchestration layer has clean surfaces to work with. That’s the prerequisite most vendors quietly skip over.
Concept deep-dive: Agentic AI orchestration
Agentic AI orchestration means software that doesn’t just recommend the next step but takes it, coordinating multiple underlying systems on a user’s behalf to complete a full workflow. Think of it as the difference between a GPS giving you directions and a self-driving car executing the route. In procurement, that means a single natural-language request can trigger supplier identification, budget validation, and RFQ issuance across separate enterprise platforms, with the human reviewing an outcome rather than managing each handoff.
Based on reporting from Procol to unveil AI procurement command centre at DPW Amsterdam, originally published 2026-09-24 02:02:00.
