The data is unambiguous. The gap between procurement leaders who have embraced AI and those still waiting is already widening and it compounds every quarter. Here’s what every CPO needs to understand about the window that is closing.
|
89%
of CPOs are racing to scale Gen AI in 2025, up from just 16% in 2024
|
3.2×
Return on AI investment for “Digital Master” CPOs vs 1.5× for followers
|
70%
Target digitalisation rate for procurement functions globally by 2027
|
£1.2M
Average annual investment in procurement digitalisation per company in 2024
|
The Conversation Has Changed
Not long ago, the question every procurement leader was being asked was: should we invest in AI? Today that question has been retired. The new question — the one sitting on every CPO’s desk in 2025 is how quickly can we close the gap with the organisations that already have.
The Hackett Group’s 2025 CPO Agenda research makes this shift stark: in 2024, just 16% of procurement leaders had prioritised Generative AI. By 2025, that figure had jumped to 89%. In a single year, AI went from a minority interest to the dominant strategic priority in the profession. That’s not a trend. That’s a pivot.
What changed? It wasn’t the technology; it had been maturing steadily for years. What changed was proof. The early movers started publishing results, and the numbers were difficult to dismiss. Procurement functions combining digital investment with talent development are now outperforming their peers across every metric that matters: cost savings, supplier performance, stakeholder satisfaction, and innovation. The performance gap is not marginal. It is structural, and it is compounding.
What the Data Actually Shows
Deloitte’s 2025 Global Chief Procurement Officer Survey, spanning over 250 CPOs across 40 countries, documents exactly what happens when procurement functions invest seriously in digital and AI capability. The results are not ambiguous.
The survey defines “Digital Masters” as organisations that combine strong technology investment with talent development. Compared to followers, these organisations report:
| Performance metric | Digital leaders | Followers |
|---|---|---|
| Cost savings — exceeded or met plan | 96% | 80% |
| Cost avoidance — exceeded or met plan | 94% | 75% |
| Internal stakeholder satisfaction | 84% | 59% |
| Supplier performance | 84% | 59% |
| Innovation enablement | 56% | 24% |
The innovation gap is particularly telling. Digital leaders are more than twice as likely to be delivering meaningful innovation through procurement. This is a function that was historically seen as a cost centre is becoming a genuine source of competitive advantage. But only for those who have built the capability to make it possible.
Dive Deeper into Agentic AI for Supplier Intelligence
Why Most CPOs Are Still Stuck at the Starting Line
If the evidence for AI in procurement is this compelling, why are so many organisations still in assessment mode? Deloitte’s survey identifies the barriers and they’re not what you might expect. This isn’t primarily a budget problem or a technology problem. It’s an understanding problem, a data problem, and a talent problem.
| # | Barrier |
|---|---|
| 1 | Internal AI and IT capability. The number one concern. Many procurement functions lack the in-house expertise to evaluate, implement, and govern AI solutions, which creates over-reliance on IT backlogs and stalled pilots. |
| 2 | Data quality. AI is only as good as the data it reasons across. Organisations with fragmented, unstandardised supplier and contract data find that AI amplifies the mess rather than resolving it. |
| 3 | Security and privacy concerns. CPOs ranked data privacy as their biggest external concern. Understandably so, given the sensitivity of supplier contracts, risk assessments, and third-party data. |
| 4 | Knowledge gap. Deloitte found that only 20% of CPO respondents have a good to extensive understanding of Generative AI, with nearly 71% describing their knowledge as limited to moderate. |
| 5 | Siloed structures. 57% of CPOs identify siloed working structures as a major obstacle to delivering value. AI can surface the insight, but if the organisation isn’t structured to act on it, the value evaporates. |
What’s striking about this list is that none of these barriers are insurmountable, and several of them dissolve entirely when AI is implemented within the right architecture. The organisations clearing these hurdles fastest are not necessarily the ones with the biggest budgets. They’re the ones who started with the right data foundation and chose implementation partners who could work within their existing environment rather than demanding infrastructure overhauls.
|
92%
Of CPOs were planning or assessing Gen AI capabilities in 2024
Deloitte CPO Survey
|
37%
Were actually piloting or deploying it. The gap is the opportunity
Deloitte CPO Survey
|
65%
Of procurement leaders are betting on AI to sharpen productivity
Gartner, 2024
|
9×
More likely to deploy advanced analytics, leading vs lagging orgs
Deloitte, 2025
|
The Three Areas Where AI Changes Everything
Not all AI applications in procurement are equal. The highest-return use cases, the ones that consistently produce results CPOs can take to the board cluster around three areas: contract intelligence, third-party risk, and supplier relationship management. These are precisely the areas where human capacity runs out first, where manual processes introduce the most error, and where the cost of getting it wrong is highest.
Consider contract intelligence alone. The average enterprise manages thousands of live contracts simultaneously. Tracking renewal dates, obligation milestones, compliance clauses, and favoured nation provisions manually is not just inefficient. It is impossible at scale. A single missed auto-renewal clause can cost anywhere from £50,000 to £500,000. An AI agent can scan an entire portfolio for that clause in five seconds. Not once. Every day, at zero marginal cost.
Third-party risk tells a similar story. Supplier risk assessments that currently consume three to four hours of analyst time can be completed in under sixty seconds when AI can query live data across multiple sources simultaneously. The analyst doesn’t disappear, they’re freed to focus on what the data means and what to do about it, rather than spending their week assembling it.
What makes agentic AI the most advanced form of the technology? Different from the chatbots most professionals have encountered is that it doesn’t wait for a question. It takes a goal, queries your live data autonomously, reasons across multiple sources, identifies patterns and anomalies, and surfaces structured insight with recommended actions. That’s not a productivity tool. That’s a capability upgrade for the entire function.
AI Doesn’t Replace Procurement Professionals – It Promotes Them.
The most persistent fear around AI in procurement that it will replace the people doing the work. It misunderstands what the technology actually does. Deloitte’s 2025 survey is explicit on this point: the organisations achieving the best results from AI are the ones treating it as a talent multiplier, not a headcount reduction strategy.
Ryan Flynn, Principal at Deloitte Consulting, put it clearly in unpacking the survey results: the goal of automation is not replacement but reduction of time spent on intensive activities like spend analysis and risk scoring, so that the humans in the function can focus on what AI cannot do. It cannot change hearts and minds. It cannot negotiate. It cannot build the trust between buyer and supplier that underpins a strategic relationship. Those remain human capabilities, and their value increases as AI handles the surrounding work.
What AI produces is capacity. The 2,500 analyst hours recovered per year per team through intelligent automation don’t disappear, they get redirected. The procurement professional who spent Tuesday pulling together a risk summary now spends Tuesday thinking through what the risk picture means strategically. That’s a better role, and it delivers more value to the business. Crucially, it’s also a more defensible role because the work being done is genuinely irreplaceable by the technology.
The Window Is Open – It Won’t Be Forever.
Every technology adoption curve in history has shown the same pattern: the leaders who moved in the early window built compounding advantages that became structural. By the time the laggards caught up, the cost of entry had risen and the gap had widened. AI in procurement is following exactly this pattern and we are, right now, inside the window where moving first still confers decisive advantage.
The Hackett Group data captures this moment precisely. In 2024, 16% prioritised Gen AI. In 2025, 89% are racing to scale it. The organisations in that first 16% have had a full year to build foundations, clean data, and run pilots. The 73% who moved from assessment to urgency in a single year are now competing for the same implementation capacity, the same talent, and the same early-mover advantages. Speed of execution is now a competitive variable in its own right.
The good news is that the barriers CPOs most commonly cite are solvable. Data quality, security architecture, and knowledge gaps are not fundamental impediments. They are implementation problems, and they have implementation solutions. The CPOs who are moving fastest are not the ones who solved every problem before starting. They are the ones who started in a controlled, secure environment with real data, and let the results build the case for the next phase.
That’s the approach worth considering: not a multi-year transformation programme, but a focused proof of value; on your data, in your environment, against the specific use cases that matter most to your function. The technology is mature enough for that. The only variable is whether you start now or later.