Three AI Signals from Ivalua’s Procurement Innovation on Tour in Dallas

The first stop on Ivalua’s Procurement Innovation on Tour 2026 took place in Dallas last week, with Chicago, Toronto, and New York to follow.

I spent the day listening to product leaders, customers, and advisors talk about AI, supplier collaboration, and the future of procurement. But the most important theme wasn’t any single feature demonstration.

It was architectural.

Across every session, one reality became clear: AI is no longer a layer that sits on top of procurement systems. It is beginning to reshape the assumptions those systems are built on.

AI Is Moving from Feature to Operating Model

For years, procurement technology has focused on workflow enablement and governance. Route the requisition. Capture the contract. Track the supplier.

In Dallas, I could hear that the conversation has shifted.

During the Innovation Showcase, Jarrod McAdoo, Product Marketing Director at Ivalua, demonstrated a multi-agent sourcing scenario where policy checks, supplier discovery, and risk screening were orchestrated in seconds before a user finalized an event. In another example, Matt Sundo, Solution Consultant at Ivalua, showed how contract summarization can extract key obligations and dates from complex agreements, turning lengthy documents into structured, actionable insight.

These tools are not just automating clicks. They are proposing actions.

Salim Khoja, Managing Director at Deloitte and lead of their Ivalua practice, described this shift as the move from workflow engines to decision engines. Instead of asking, “How do we automate the process?” the question becomes, “How do we delegate parts of the process under defined guardrails?”

That is a fundamentally different operating model.

Rex Roedger, VP of Supply Chain at General Atomics Aeronautical Systems, put it plainly. AI, in his view, is not about replacing buyers. It is about removing the “boring, mundane” work of assembling data and drafting standard documents so that procurement professionals can focus on strategy and negotiation.

In other words, AI is beginning to change not just what the system does, but what the human does.

That shift only works if the underlying architecture supports it.

Architecture Is Becoming Strategic

One of the most consistent themes from the customer panels was the importance of foundational design decisions.

Emily Roth, Manager of Digital Transformation at Texas Instruments, spoke candidly to me about building a single supplier core before expanding broader procurement capabilities. The goal was not simply consolidation. It was to create one place where supplier data could be trusted and reused across workflows.

Srinath “Sri” Iyengar, Director of IT Solutions Management for Global Supply Chain at Applied Materials, described replacing a 15-year legacy collaboration platform. The challenge was not feature parity. It was regaining configurability without hard-coding processes into a system that could not evolve. Over several years, Applied digitized sub-tier relationships for thousands of critical parts, enabling what Sri referred to as “collision point” analysis across shared suppliers.

Stephanie Cenal, also in IT Solutions Management at Applied Materials, reinforced that this was not a lift-and-shift exercise. It was an opportunity to level up supplier collaboration, standardize processes, and create a supplier console that guides partners toward the most important daily actions.

General Atomics, at an earlier stage of its journey, emphasized scalability and integration over narrow optimization. Rex Roedger was clear that compliance in defense is non-negotiable. But compliance supported by spreadsheets and disconnected tools does not scale. His team is betting on a collaboration environment that can tie together supplier onboarding, documentation, and chain-of-custody requirements in a way that reduces manual reconstruction during audits.

None of these stories were framed as platform-versus-point-solution debates. They were framed as architecture questions.

Salim Khoja shared research from Deloitte’s CPO survey distinguishing digital masters from digital followers. The difference was not who had more tools. It was who had coherent data and integrated processes. Organizations with unified or tightly orchestrated data models consistently see materially higher returns on digital investments.

And AI will only widen that gap.

If your supplier, contract, risk, and sourcing data do not align, AI will spend most of its effort reconciling inconsistencies. If your architecture is coherent, AI compounds value.

AI amplifies whatever foundation you give it.

What This Means for Procurement Leaders

If you are building a technology roadmap for 2026 and beyond, three implications stand out.

First, data hygiene is no longer a background project. It is a strategic enabler. Turning on AI in a fragmented environment will expose weaknesses quickly.

Second, integration discipline matters more than feature count. The marginal value of the next capability is lower than the value of making existing capabilities work together seamlessly.

Third, AI roadmaps must align with operating model design. Delegating tasks to agents requires clarity about policy, authority, and human oversight. Technology will not fix ambiguous governance.

The next three stops on the Procurement Innovation on Tour series in Chicago (May 19), New York (September 15), and Toronto (October 6) will no doubt surface additional perspectives - and we will be there to cover them all.

The signal from Dallas was very clear. The AI conversation in procurement is maturing.

The real competitive advantage will not come from turning on the newest capability first. It will come from making deliberate architectural choices that allow those capabilities to operate at scale.

In 2026, AI forces clarity about data ownership, forces discipline in integration, and forces leaders to define what should be delegated and what must remain human.

Architecture used to be a background decision. Increasingly, it is now a strategic one.