Agentic AI and Orchestration Are Key to Transforming Procurement, Suggests Recent Study

If you work in procurement or supply chain, you already know the feeling. The pressure to do more with the same resources. The nagging sense that your data is not quite trustworthy enough to bet a strategy on. The technology landscape shifting faster than your team can keep up with. And now, on top of all that, everyone wants to talk about AI.

The Hackett Group recently sat down with procurement leaders across major global organizations to get a clear picture of where the function is heading. What came back was honest, grounded, and in many ways reassuring: the challenges you are facing are widely shared, and the path forward is becoming clearer.

Here is what stood out.

How it Works The Priorities Are Shifting Fast — But Not Evenly

When leaders were asked where they expect the greatest near-term changes, information and analytics came out on top, followed closely by technology solutions and skills and talent. The organizational model and outsourcing arrangements, by contrast, are expected to change far less dramatically.

What that tells you is that procurement leaders are not planning a wholesale structural overhaul. They are investing in smarter tools, better data, and sharper people. The bones of the function are staying put. What is changing is what it runs on.

This matters because it means the ROI conversation is shifting from "should we restructure procurement?" to "are we getting the most out of what we already have?"

The Honest Truth About Data

Here is the finding that keeps coming up regardless of the topic: data quality is the bottleneck for almost everything else.

Leaders want better AI. They want stronger category strategies. They want real-time risk insights. And across the board, the barrier is the same. Poor spend classification, inconsistent supplier records, and unstructured contract data are slowing progress in supplier management, analytics, and contracting alike.

Analytics and AI simply cannot deliver value without clean, trusted data. This is not a new observation, but it keeps being the most important one. If your organization is planning significant technology investments this year, the highest-leverage thing you can do first is fix the foundations.

Agentic AI Is Coming In Hot — With Appropriate Caution

Nobody surveyed said they have no plans for agentic AI. Zero percent. That is a striking number.

The majority are still in evaluation or cautious piloting mode, which is completely appropriate given where the technology sits today. But the direction is unmistakable. And the expectations are high: 89% of organizations surveyed expect agentic AI to improve process efficiency and autonomy, more than half expect better insights and decision support, and around half expect improvements in data quality and user experience.

The processes where AI maturity is already highest tend to be the structured, repeatable ones: procure to pay, spend analytics, and parts of sourcing. Intake management and risk monitoring are earlier stage, largely because fragmented systems and disconnected data make it harder for AI to get traction.

The practical takeaway: start where the data is cleanest and the workflows are most consistent. Build outward from there

Orchestration Is the Word Everyone Keeps Using

One of the most interesting themes in the research is how much attention procurement leaders are giving to orchestration. Not just AI. Not just automation. The coordinated design and management of procurement activities, systems, data, and stakeholders into a unified process.

78% of leaders view orchestration as highly or extensively important to their overall strategy. And it is easy to see why.

The problem many organizations are trying to solve is not a lack of tools. It is a surplus of them, disconnected from each other and creating confusion for the people trying to use them. Stakeholders do not know where to start a request. Requests get routed inconsistently. Approvals pile up. Spend leaks out through informal channels.

Orchestration is the answer to that fragmentation. A unified front door. Standard workflows. Clearer routing. Real-time visibility into where things stand. And crucially, the structured foundation that AI needs to actually automate steps rather than just augmenting human effort at the margins.

Risk Management Is Reactive for Most Organizations — And That Is a Problem

Here is a sobering statistic: the median organization has proactive risk monitoring in place for only 43% of its purchasing categories. Top quartile performers reach 63%. Which means even the best performers are leaving more than a third of their categories without early warning systems.

The risk environment in 2026 is not forgiving. The top five enterprise risk concerns identified in the research are cybersecurity, trade wars, economic downturn, the impact of AI-enabled technology, and regulatory compliance. What procurement leaders want is real-time alerts, automated monitoring, and predictive insights. What they currently have in most cases is manual processes and reactive responses.

The path is well understood: build better dashboards, integrate external risk feeds, harmonize supplier master data, and embed risk directly into category strategies rather than treating it as a separate workstream.

Category Management Is Growing Up

The structural side of category management is actually in pretty good shape. Most organizations report well-defined strategies, good stakeholder partnerships, and a reasonably scalable operating model.

Where there is room to grow is in data integration. Only about half of organizations feel their market intelligence, risk data, and compliance information are well integrated into category planning. The vision is clear: strategies driven by real market dynamics and supplier health data rather than historical spend patterns alone.

AI is starting to play a role here, accelerating the analysis of market trends and cost drivers. But the human judgment, the commercial instincts, the stakeholder relationships — those remain squarely with the category manager. AI in procurement is a support function right now, not a decision-maker.

What This All Adds Up To

The leaders who are pulling ahead are not necessarily the ones with the most advanced technology. They are the ones who have done the disciplined work of cleaning their data, simplifying their processes, building their teams' capabilities, and making it easy for stakeholders to engage.

That is not glamorous work. But it is the work that makes everything else possible.

This blog draws on findings from The Hackett Group's 2026 Procurement Executive Insight Report, developed in partnership with GEP.

Download the full report.