The hype around AI agents in procurement is justified. They’re fast, efficient, and increasingly capable of performing complex tasks. But here’s the uncomfortable truth: AI agents, left on their own, won’t transform your procurement function. What will? Orchestration.
Agentic AI—the concept of semi-autonomous systems that take action on behalf of users—is reshaping procurement. But without orchestration to connect data, people, and processes across fragmented tech stacks, these agents risk becoming just another disconnected tool in an already cluttered landscape.
So if your organization is already exploring GenAI, it’s time to move beyond experimentation and think strategically—not just about what AI agents can do, but about how orchestration enables them to deliver value at scale.
AI agents can read contracts, route requests, and surface relevant documents from vast data repositories. They excel at automating routine tasks and speeding up operations—think summarizing supplier proposals or extracting invoice line items. But when it comes to making context-aware decisions that cut across systems and teams? That’s where orchestration steps in.
Think of AI agents as brilliant interns. They’re smart, eager to help, and good at handling repetitive work. But without a manager—or in this case, a system to guide, escalate, and coordinate their tasks—they struggle to deliver outcomes that require judgment, sequencing, or collaboration.
Orchestration platforms do what AI agents can’t do alone: they unify processes across disparate tools, teams, and decision points. They handle the “glue work” of procurement—like coordinating supplier onboarding with risk assessments and legal approvals, or triggering contract workflows based on spend thresholds—all without manual handoffs or swivel-chair integrations.
This isn’t just about automation. It’s about enabling intelligent, end-to-end experiences. In a well-orchestrated environment, an AI agent isn’t just extracting data—it’s triggering the next best action in a broader workflow, all while keeping stakeholders aligned and compliant.
At GSK, orchestration became the key to unlocking a better user experience. Faced with inconsistent procurement paths and disjointed systems, they layered orchestration on top of their existing tools to guide users through the process—showing them what to do, when to do it, and why it matters. The result? A cleaner, more compliant, and far more agile way to work.
To deploy agentic AI successfully in procurement, organizations need three things:
Pierre Mitchell of Spend Matters puts it well: “Procurement isn’t broken—it’s disconnected.” Orchestration reconnects the dots. And when you layer in agentic AI, you create a system that’s not only smarter, but truly transformative.
While AI agents today act more like co-pilots—assisting users, offering recommendations, and automating sub-tasks—the trajectory is clear. We’re moving toward multi-agent environments where agents collaborate, learn, and manage complete procurement workflows with minimal human input.
But again, the key enabler isn’t the AI itself—it’s the orchestration framework around it. Without it, agents don’t know where to go next, which systems to update, or how to escalate a risk. With orchestration, those agents become exponentially more valuable.
As Art Clark (former Apple AI leader) noted in a recent webinar with Spend Matters, the most successful organizations aren’t waiting for perfect AI. They’re deploying agentic systems today for tasks where correctness can be easily verified—and orchestrating them into broader business processes that are already well understood.
Procurement leaders should follow that lead: start where confidence is high, orchestrate for impact, and grow from there.
In this new chapter of digital procurement, agents may be the muscle—but orchestration is the brain.