For decades, MS Excel has quietly powered the procurement function in many capacities, both upstream and downstream. It’s flexible, familiar, and—critically—under user control. Even after multiple waves of enterprise tech, teams still default to spreadsheets and email because most systems feel too slow, rigid, or disconnected from how procurement really works.
Now, a new force is accelerating that same do-it-yourself mindset: generative and agentic AI. With tools like OpenAI’s ChatGPT, Anthropic’s Claude, and Google Gemini, procurement professionals can spin up RFP drafts, generate supplier scorecards, summarize market research, or even build mini decision engines—instantly. The friction is gone. The result? A surge in homegrown solutions that look a lot like the Excel shadow systems of the past—but move faster, proliferate wider, and carry even more risk.
Why This Is Familiar—yet Different
AI is unlocking exactly what Excel once offered: speed, autonomy, and flexibility. Need to build a new supplier intake workflow? An AI tool can mock up the first version. Want to summarize contract terms? A prompt can extract key clauses. What’s different now is personalized access to AI at a new level of personalized scale. This next generation of AI in the form of generative and agentic tools make it far easier for every team member, not just the Excel superusers, to create bespoke workflows in minutes.
In the short term, this looks like progress. Teams seem more productive. Ideas get tested more quickly. Processes feel more responsive. But beneath the surface, the same old cracks are starting to show.
Silos, Misinformation, and the Loss of Control
Just like Excel once created pockets of hidden data, these next gen AI tools can quickly fracture the procurement landscape. Prompt outputs live on desktops or in chat windows. Decisions are influenced by models no one can audit. Processes evolve in parallel without coordination. There’s no common language, no shared data foundation, no traceability.
This isn’t just a governance problem. It’s a performance problem. When sourcing decisions are based on local models or potentially hallucinated insights, the organization starts making the wrong bets. And because no one sees the full picture, the errors compound.
The Real Risk Isn’t Security. It’s Suboptimal Procurement
Yes, shadow AI poses security concerns. But those will be addressed—through enterprise controls, policy frameworks, and better tooling. The deeper risk is strategic misalignment. If every team builds their own agentic assistant to handle sourcing, intake, or supplier scoring—without shared logic or data—they’re solving today’s pain at the cost of tomorrow’s cohesion.
Left unchecked, this trajectory leads to an even messier version of the Excel era: fragmented decision logic, duplicated tools, disconnected insights. Procurement gets faster, but not smarter or more efficient.
There’s a Better Path: Reinvent the Operating Model
Generative and agentic AI are not just a faster Excel. It’s a chance to redesign how procurement operates. That means shifting from DIY experimentation toward intentional architecture. From individual agents to coordinated intelligence. From workaround tooling to strategic capabilities.
And that doesn’t mean reverting to rigid systems. A new class of tools is emerging—native AI platforms designed for procurement. These blend generative and agentic capabilities to support tasks like always on intelligence, autonomous sourcing, intelligent Intake and Orchestration (I&O), and dynamic supplier engagement. These aren’t generic copilots—they’re purpose-built systems that can learn, adapt, and connect across the procurement lifecycle.
The opportunity is to move beyond task-level acceleration toward true workflow reinvention. To harness AI not just to replicate the old ways faster, but to architect smarter, more transparent, and value-aligned processes from the ground up.
It starts with a clear mandate: channel the creativity next gen AI enables into a shared system of value. Build an AI operating layer that’s governed, data-rich, and procurement-specific. Let teams create—but on rails. Make it easy to innovate without going rogue.
Done right, next gen AI won’t just automate current workflows. It will change what’s possible. And procurement won’t just survive the next shadow wave—it’ll lead it.
At Liberis we work with startups and growth-stage procurement and supply chain solution providers to clarify and amplify their voice so they can stand out, resonate with enterprise buyers, and scale effectively.
👉 Learn more at Liberis Consulting.
