
Next-gen supply chains
Resolving the AI autonomy dilemma
While procurement leaders envision autonomous supply chains where AI manages negotiations and decisions with minimal oversight, the reality remains uneven. Firms across the US and Europe are layering AI onto fragmented, legacy systems. Consequently, AI succeeds in structured, rules-based tasks like invoice processing and spend categorisation, while complex work involving relationships and judgment remains human-led. Beyond technical integration, procurement traditionally lacks data engineering and AI governance capabilities. Few firms possess the skills to manage autonomous systems, and employees worry automation will disrupt traditional structures and entry-level talent development.
Successful firms share key approaches: investing in data architecture first, redesigning roles alongside technology, and treating human oversight and governance as requirements for scaling AI.
Research highlights 2026
Start of carousel.
01
Rhetoric v reality: boards talk of agentic autonomy, but AI still runs on legacy software and incomplete data
There is a widening gap between executives' vision for artificial intelligence (AI) and the reality of deploying it. Across firms in America and western Europe, companies are bolting AI onto legacy procurement and supply-chain systems without rebuilding the data architecture needed to give it real-time access to information across the business. As a result, AI rarely runs procurement processes from end to end. Instead, it automates isolated tasks that depend on narrow data sets. Firms that want AI to deliver lasting gains should focus less on superficial pilots and more on building the data architecture that makes autonomous systems possible.
02
The investment gap: firms acknowledge their data challenge but few invest to solve it
While 91% of firms acknowledge the need for greater investment in data engineering to enable autonomous supply chains, and 85% say their software stack needs modernising, few plan to increase spending meaningfully next year. Most expect to spend either the same amount or only slightly more. That mismatch between ambition and investment threatens to undermine their AI strategies. Firms that treat data engineering as a marginal IT expense rather than a strategic investment are unlikely to build the foundations that autonomous systems require.
03
Compliance before creativity: AI's insights will be ignored unless they can be audited, tracked and explained
AI agents can process far more data than any human and identify patterns that would otherwise go unnoticed. That promises large productivity gains, but it also creates a problem. Procurement managers are reluctant to act on recommendations they cannot explain, justify or audit, and regulators may refuse to accept them altogether. As a result, firms grant AI far less autonomy than the technology appears to permit. That is unlikely to change until AI systems are designed with governance in mind, embedding clear explanations and audit trails into their decisions.
04
The skills mismatch: procurement needs AI engineers who understand the business, not just negotiators
Procurement has long rewarded commercial instinct and negotiating skill. AI-enabled supply chains demand something different: data and engineering expertise. The result is a widening skills gap. Executives report a surplus of commercial talent but a shortage of the technical skills they now value most. Future procurement teams will need two forms of expertise: an understanding of AI, data and software, and the commercial judgement to identify tasks that can be automated. The first remains scarce.
Human judgement, by contrast, is plentiful. Two-thirds of firms report having it, yet they rank it well below technical expertise. Leading firms are creating roles such as AI compliance auditors and AI governance specialists, often recruiting from outside. Those that fail to build these capabilities internally will struggle to deploy, oversee and improve the AI systems on which they increasingly depend.
05
The entry-level gap: automation is eliminating the training ground for tomorrow's talent
Procurement has traditionally trained future leaders by giving junior staff routine work: tracking purchase orders, processing invoices and categorising spend. Yet these are precisely the tasks AI is automating fastest. At most firms they are already largely automated.¹ As procurement becomes more autonomous, it risks closing off the apprenticeship through which commercial judgement has long been developed.
Firms will therefore need to rethink how they train procurement professionals. Machines may increasingly execute routine work, but people will still need to understand how procurement functions, alongside the data, software and AI systems that now support it. The challenge is not simply to teach technical skills. It is to replace the experience that automation is removing.
1 See Briefing Paper One: "The Agentic AI Implementation Challenge"
06
Beyond efficiency: AI's next role is discovery
So far, most firms have used AI to perform familiar procurement tasks more efficiently, such as chasing invoices. The most innovative are using it to do things that were previously impossible: identifying alternative materials or testing supply chains against scenarios that would take humans weeks to model. In doing so, they are shifting procurement from an operational function to a source of strategic advantage. Firms that use AI only to cut costs will achieve incremental gains. Those that use it to generate new insights may redefine how they compete.
End of carousel.
Latest content
Start of carousel.

Agentic should-cost modelling is changing procurement and product design

Reskilling procurement teams for the age of agentic AI

More than the ghost in the machine: human agency in agentic AI deployment

The Agentic AI Implementation Challenge

Agent of agents: working towards total agentic orchestration in procurement

How agentic AI is redesigning procurement and supply-chain operations

Building autonomous procurement and supply chains: the systems that make it work
End of carousel.
Insights from GEP

Agentic AI Buyer’s Guide
How To Evaluate Enterprise-Ready Agentic AI for Procurement and Supply Chain
Start of carousel.

AI-Native vs. Bolt-On Procurement Platforms

Why Domain Expertise Is an Essential Capability for Agentic Procurement Orchestration

Is Your Procurement Intake and Orchestration Platform Really Intelligent?

The Supply Chain AI Readiness Report

The AI-Ready Procurement Team: Upskilling for the Next Era of Value Creation
End of carousel.