I started my career in 2003 as an e-sourcing analyst. The work was largely transactional — running RFQs, maintaining supplier data, processing requisitions. Useful work. Necessary work. Also the kind of work that, if I'm being honest, a well-configured system could have done better than I did on most days.
That was twenty-two years ago. The tools have changed dramatically. The underlying question hasn't: where does human judgement add value that a system cannot replicate?
AI is sharpening that question considerably. And the procurement professionals who will thrive in the next decade are the ones who have a clear answer to it.
The work that AI is taking first
Let's be precise about this. AI isn't taking procurement jobs wholesale. It's taking specific tasks — the ones that are high volume, rule-based, data-dependent, and repeatable.
Invoice matching. Purchase order routing. Spend classification. Supplier onboarding checks. Contract clause extraction. Savings tracking against baseline. These are tasks that have occupied significant portions of procurement team capacity for decades. They're also tasks where human involvement adds relatively little value once the rules and data are well-structured.
If your role is primarily defined by tasks in that category, the honest conversation is not "will AI affect me?" but "what am I building towards instead?"
What moves up the curve
The work that AI handles badly — at least for now, and likely for longer than the vendor conversations suggest — is the work that requires genuine contextual judgement.
Knowing when a technically compliant supplier recommendation is commercially wrong because of a relationship dynamic that doesn't appear in the data. Understanding why a stakeholder who is saying yes in the room is actually saying no in their behaviour. Designing a category strategy that works within the political constraints of an organisation as well as its commercial ones. Deciding when the numbers say one thing and the right answer is something else.
These are not tasks that get automated. They get more valuable as the routine work disappears around them.
The procurement professionals I see thriving in the next cycle share a few characteristics. They've invested in commercial literacy — they understand what the business is trying to achieve, not just what procurement is trying to buy. They've built real stakeholder relationships, not just stakeholder management processes. They can read a data output critically — understanding what it means, what it's missing, and when to trust it versus when to question it.
And increasingly, they understand AI well enough to govern it — to know when an agent's recommendation needs human review, and why.
The skills that compound
The question I get asked most often when I'm mentoring procurement professionals is some version of "should I be worried about AI?" My honest answer is: only if you're not moving.
The professionals who will be displaced are the ones who treat their current role as a destination rather than a point on a curve. The ones who are building — commercial acumen, stakeholder influence, data literacy, AI fluency, genuine industry expertise — are the ones for whom AI becomes a tool rather than a threat.
This isn't about becoming a technology person. It's about being a procurement person who understands technology well enough to work with it, govern it, and occasionally push back on it when it's wrong.
What this means if you're early in your career
If you're five years in and wondering what to build towards — the answer is deliberate. Don't just get good at your current role. Understand the layer above it. If you're running sourcing events, understand the category strategy behind them. If you're managing supplier data, understand the commercial relationships that data is supposed to serve. If you're working in P2P operations, understand what the business is trying to buy and why.
Every layer of context you add makes you harder to replace and more valuable in the conversations that matter.
The curve is real. The opportunity is to be further up it than your job title currently suggests.
