GEP
When AI Starts Thinking Like Procurement Professionals

Procurement software has long promised efficiency, but often delivered complexity, with more data, steps and still too little intelligence.

Now GEP, a leading global provider of procurement software, strategy and managed services, believes a new generation of AI can change that. Built from decades of procurement experience, its emerging agentic AI systems are being trained to perceive, reason and act in context, to think less like code and more like a procurement professional.

The phrase AI-first has become corporate wallpaper, an easy claim often detached from reality. In procurement, especially, the term appears on nearly every software vendor’s website, yet most "AI-powered" tools still deliver modest results. They automate tasks but miss the real complexity: the thousands of micro-decisions, trade-offs and judgment calls that define how organizations actually buy.

“Procurement isn’t a mechanical process you can just plug generic AI into,” says Santosh Nair, global head of growth, software, at GEP. “It’s a living ecosystem. Every sourcing decision touches suppliers, finance, risk and sustainability. If AI doesn’t understand that web of relationships, it adds noise, not intelligence.”

Building from the Ground Up

GEP has spent the past several years pursuing a simple idea: that AI must be built for procurement, not merely applied to it. The company’s engineers have been designing systems that learn from 25 years of real sourcing data — contracts, supplier records, market indices and performance histories — rather than the generic business datasets that fuel most enterprise AI.

The result is GEP SMART, a next-generation, agentic AI-powered procurement software conceived to operate end-to-end across the source-to-pay process. The platform is not designed to automate tasks in isolation but to orchestrate them.

“We built it as a native AI architecture,” says Nair. “There’s no bolt-on intelligence here. Every function, from spend analysis to supplier management, was designed to learn, reason and act with context.”

The Coming Age of Agentic AI

While today’s AI systems largely execute tasks when prompted, the next wave, known as agentic AI, is expected to act with greater autonomy. Instead of waiting for instructions, these systems will understand intent, anticipate needs and coordinate actions across workflows.

Procurement isn’t a mechanical process you can just plug generic AI into. It’s a living ecosystem. Every sourcing decision touches suppliers, finance, risk and sustainability. If AI doesn’t understand that web of relationships, it adds noise, not intelligence.

GEP SMART was built with this future in mind. The company envisions multiple specialized AI agents, each focusing on sourcing, compliance or risk, working in parallel and coordinated by a super-agent that ensures alignment with policy and strategy.

The possibilities are striking.

In the near future, such systems could detect early signs of supplier distress, suggest contingency plans or even propose revised contract terms based on shifting market conditions, all while giving human teams the final word.

“We’re building for what’s next,” says Nair. “A world where AI acts like a trusted team member, not just a faster assistant.”

Intelligence with Depth

The intelligence behind GEP’s software draws from a wide spectrum of information. Traditional enterprise systems rely on structured data from ERPs. GEP’s platform fuses that with unstructured sources such as emails, transcripts, ESG reports and regulatory updates to build a complete picture of procurement.

That multimodal approach is what Nair believes separates practical AI from performative AI.

“Procurement decisions are made in context,” he says. “You can’t predict supplier risk from spreadsheets alone. You need the contract terms, the performance history, the compliance notes, all of it, to make a meaningful recommendation.”

By linking this data to a connected layer of supplier profiles, benchmarks and historical outcomes, GEP’s AI can identify patterns that humans might overlook. It does not yet execute those insights autonomously, but that is where procurement appears to be heading.

Designing for Flexibility

A key part of GEP’s strategy is flexibility. The platform is LLM-agnostic, meaning enterprises can choose the large language model that best fits their needs or regulations. It is also cloud-neutral, giving companies control over where their data resides.

Equally important, GEP has embedded low-code and no-code tools so users can adapt workflows without waiting for IT intervention. Nair believes this responsiveness will define the next generation of procurement software.

“Markets are changing weekly. The technology can’t take months to catch up. If AI can’t adapt quickly, it’s not really intelligent,” he says.

The Human Role Evolves

Even as automation grows more capable, GEP envisions humans and AI working in concert. The goal is not to make procurement more human-centric or more automated, but more intelligent. AI handles the scale and speed of execution, while people apply judgment and creativity to the decisions that matter most.

In the near term, this means a division of labor. AI will handle the repetitive, time-sensitive groundwork such as risk scans, supplier discovery and compliance checks, while people focus on negotiation, planning and stakeholder management. Over time, as confidence in agentic systems grows, those boundaries may blur.

“The endgame isn’t full automation,” Nair says. “It’s collaboration. Humans will still make the calls. AI will just make sure they have better information, faster.”

The Road Ahead

The shift to agentic AI is still in its early stages. Gartner predicts that by 2028, one in three enterprise applications will feature some level of autonomous reasoning. For procurement, the change could be transformative. Imagine category strategies that evolve in real time, or supplier contracts that adapt automatically to policy changes. Imagine risk management systems that detect disruptions before they happen. These are possibilities, not promises, but they are moving quickly from concept to capability.

GEP’s ambition is to lead that evolution.

“Procurement has always been about balancing competing priorities: cost, compliance, sustainability, innovation,” Nair says. “AI won’t eliminate that complexity. It will help us navigate it.”

For GEP, the future of agentic AI is not a distant concept but an emerging reality. The groundwork has already been laid, and early deployments are proving what’s possible. Procurement organizations are beginning to see how intelligence, once theoretical, can now act, reason and guide decisions in real time. The shift is already underway.