McKinsey: How AI Can Unlock Value for Procurement

Procurement is under pressure like never before. Rising tariffs, inflation and persistent supply disruptions are forcing organisations to rethink long-standing processes. Budgets are tight, talent is scarce and expectations for savings continue to climb.

Against this backdrop, artificial intelligence is emerging as a critical enabler, transforming procurement from a tactical back-office function into a strategic nerve center.

From automating workflows to creating entirely new roles and capabilities, AI promises significant productivity gains and competitive advantage. McKinsey Partner Samir Khushalani and Expert Engagement Manager, Dan Albrecht, spoke to Procurement Magazine to explore how organisations can act fast – or risk falling behind.

Why is procurement under more pressure today than at any point in the past?

Samir Khushalani:
Procurement is operating in a world where volatility, complexity and uncertainty are no longer episodic – they’re constant. Trade protectionism and economic nationalism are accelerating, with tariffs increasingly replacing trade agreements.

In the US alone, the trade-weighted average tariff rate rose from under 2% in 2024 to 17% by October 2025, forcing companies to re-examine their supply base and long-established trade corridors.

What additional operational and resource pressures are procurement teams facing?

Dan Albrecht:
On top of macro disruption, procurement teams are being asked to deliver more with less.

In a recent McKinsey survey, 55% of procurement leaders reported flat or shrinking budgets, even as every respondent said their savings targets had increased.

Spend managed per full-time employee is now about 50% higher than it was five years ago. At the same time, procurement still struggles to attract top talent, partly because it’s too often viewed as a tactical cost-cutting function rather than a strategic partner.

Why do you believe incremental change is no longer enough?

Samir Khushalani:
The scale and persistence of today’s challenges demand a fundamental reimagining of the function. Incremental improvements won’t keep pace. Artificial intelligence is emerging as the critical enabler that allows procurement to operate at a completely different level.

The question is no longer whether AI will transform procurement – it’s how quickly organisations will move, and who will lead.

How far along is procurement in adopting AI today?

Dan Albrecht:
Many organisations are experimenting, but most are still early in the journey. In our survey of more than 300 global procurement leaders, 40% said they are actively piloting generative AI tools. Some are already seeing tangible results.

For example, one pharmaceutical company piloted an AI-based invoice-to-contract reconciliation tool to uncover more than US$10m in value leakage. These early wins are just the beginning and need to be scaled to achieve the full potential.

What kind of productivity impact can AI deliver in procurement?

Samir Khushalani:
McKinsey estimates that AI copilots, chatbots and task-level tools can improve procurement productivity by 25 to 40%. But the real leap forward comes when AI moves beyond individual use cases.

The next frontier is agentic procurement – AI agents that autonomously execute end-to-end workflows designed specifically for machine-led execution.

Over time, this scales into agentic factories where multiple agents collaborate with each other and access enterprise systems, with humans serving as the final line of accountability.

How does that shift change procurement’s role in the enterprise?

Samir Khushalani:
In this future, procurement becomes an enterprise “nerve center.” The function will be smaller and flatter, yet far more influential.

AI-native workflows across source-to-pay will increasingly run autonomously, pushing the marginal cost of procurement toward the cost of compute. Human effort will shift from executing transactions to orchestrating business outcomes that truly matter.

How will success be measured in an AI-enabled procurement model?

Dan Albrecht:
Performance metrics will fundamentally change. Process-centric measures, such as purchase-order cycle time, will give way to outcome-oriented ones.

Instead of asking how long it takes to create a PO, leaders will ask: what percentage of POs are created autonomously – and with an error rate below 1%? That’s a very different way of thinking about value.

What happens to traditional procurement roles as AI becomes embedded in daily work?

Dan Albrecht:
Roles will evolve quickly. Agentic systems will take on much of the data analysis and routine execution, while humans focus on work requiring judgment and emotional intelligence.

Buyers and accounts payable specialists may oversee AI outputs or act as escalation points. Category managers will spend more time deeply understanding business needs and strengthening relationships. In fact, 43% of CPOs in a recent McKinsey survey identified strategic thinking as the most critical future competency for category managers.

What new roles will emerge as a result of agentic procurement?

Samir Khushalani:
We’ll see entirely new roles appear. Procurement AI strategists will set priorities for AI deployment and drive adoption across the workforce.

AI risk and ethics specialists will ensure the responsible and transparent use of AI, including avoiding unintended bias in supplier selection. And AI orchestrators will manage the growing constellation of agents, preventing uncontrolled AI sprawl.

Where will these new capabilities typically sit within the organisation?

Dan Albrecht:
In many cases, they’ll be anchored in a revitalised procurement centre of excellence. Existing COEs often provide a strong foundation, but most organisations aren’t yet set up for success.

Our research shows that only about one in ten organisations offers the full range of services required to unlock a COE’s potential. Those that do see outsized returns.

Can you share an example of what a high-performing COE can deliver?

Dan Albrecht:
One specialty chemicals company used its COE to develop best-in-class should-cost models grounded in materials, labour and market data. That capability alone delivered 13% savings on raw materials. It’s a powerful example of how advanced analytics and AI can translate directly into value.

How will AI-optimised COEs differ from today’s models?

Dan Albrecht:
The COEs of the future will have a broader remit and higher profile. They’ll provide deep expertise in AI-augmented cost engineering, next-generation analytics, AI-native process excellence, supplier risk management and capability building– all while remaining tightly integrated with day-to-day procurement execution.

Why is the urgency to act so high right now?

Samir Khushalani: 

Procurement is falling behind other functions. AI adoption in marketing and sales is roughly six times higher than in procurement. That means suppliers are moving faster than buyers to gain competitive advantage. The cost of waiting is rising, and the window to lead is narrowing.

What’s the takeaway for procurement leaders today?

Samir Khushalani:
We are at a rare inlection point for procurement. The procurement leaders of tomorrow will be the first movers today. Those who wait will find the gap increasingly difficult to close. There is no shortage of ways that AI can be used to disrupt Procurement. The question is which areas should a company prioritise based on the impact at stake.

The AI driven transformation of the procurement function must be accompanied by an equal if not greater focus on the people – changing mindsets and behaviours, increasing comfort with and trust in agentic systems and equipping practitioners with the skills needed to perform their future roles.