For most of my career, the senior procurement leader's job was defined by what they managed: spend portfolios, supplier relationships, transformation programmes. The value was in the outcomes — savings delivered, risks mitigated, processes redesigned. Leadership was largely a function of doing the work well enough, at sufficient scale, that others wanted to follow your approach.
That definition is changing. Not because the outcomes matter less, but because AI is now handling an increasing share of the execution — and what's left, increasingly, is the work that only humans can do. Specifically: deciding what matters, building the relationships that make strategy executable, and developing the people who will carry the function forward.
The organisations getting this transition right are the ones where senior leaders have updated their answer to a basic question: what is my primary job right now? If the answer is still "managing the category portfolio" or "running the transformation programme," that may be accurate today. In three years, it will be incomplete.
What AI changes about leadership
The shift isn't dramatic on the surface. Day-to-day procurement work continues. The categories still need managing. The transformation still needs leading. But underneath the surface, the composition of value is shifting.
Hard skills — the ability to manage complex spend analysis, maintain technical compliance data, execute sourcing events to a reliable methodology — are being commoditised faster than most people in the function are acknowledging. When any organisation can buy an AI tool that does spend analysis reliably and at scale, spend analysis stops being a competitive differentiator. It becomes table stakes.
What doesn't get commoditised is judgement built on context. The ability to look at a commercially sensitive supplier situation and know which levers to pull. The stakeholder relationship that makes a difficult conversation possible rather than destructive. The instinct — developed through years of exposure to how organisations actually work — that tells you when the process is right and the answer is still wrong.
These capabilities don't emerge from training courses. They emerge from deliberate development over time, usually through exposure to senior thinking and honest feedback. Which means someone has to provide that. At scale. Intentionally.
Why mentorship has to become a discipline
The word mentorship gets used loosely in organisations. It often means an occasional conversation with someone more senior, usually at the mentee's initiative, with no particular structure and no accountability on either side. That's better than nothing. It is not what this moment requires.
Moving a procurement team from First Curve transactional work to Second Curve strategic value — the transition I've written about in the context of career development and the human-first imperative — requires deliberate, structured intervention. You cannot tell someone to be more strategic and expect the capability to appear. You have to build it, over time, through intentional exposure and honest feedback.
Three things matter most in that process.
Teaching commercial context, not technical process. When routine tasks are automated, the question your team needs to be able to answer shifts. It's no longer "how do I run this sourcing event?" It's "why are we running this sourcing event, what are we trying to achieve, and how does this supplier relationship serve the business's actual goals?" That's a different kind of thinking, and it develops through exposure to the conversations where those questions get asked — not through process training.
Building psychological safety around AI outputs. In organisations where AI tools have become authoritative — where the system's recommendation is treated as the answer rather than a hypothesis — you create a specific kind of risk. The model is wrong in ways that an experienced practitioner would catch. But if the culture treats challenging the AI output as friction or excess caution, nobody catches it. The leader's job is to make it normal, even expected, for team members to question what the system produces. That starts with modelling it yourself.
Investing in relationship equity before it's needed. The most valuable supplier relationships are not built in a crisis. They're the ones that were maintained consistently during the periods when everything was working — so that when something goes wrong, there's a foundation of trust to draw on. Teaching your team to build and maintain those relationships, as a practice rather than an exception, is one of the highest-value investments a senior leader can make. It shows up in resilience rather than efficiency metrics, which is probably why it's chronically underfunded.
The misuse of talented people
I want to be direct about something that doesn't get said often enough in these conversations. If an AI tool can straightforwardly replace someone in your team's current role, the problem that reveals is not primarily a technology problem. It's a question about how you've been developing that person.
The most expensive thing an organisation can do with a commercially capable procurement professional is keep them processing transactions and maintaining data. It's not just the cost of their time doing work the system could do better. It's the opportunity cost of the strategic contribution they're not making — the supplier relationships not built, the category strategies not developed, the stakeholder conversations not had.
Automation clears the desk. What's on the desk next is a leadership decision. Make it deliberately.
The function that emerges on the other side of this transition — leaner on headcount, significantly higher on impact — will be built by leaders who treated people development not as an HR obligation but as a strategic investment. The teams that will matter in five years are being developed now, in the conversations and opportunities that senior leaders choose to create or not create today.
Step away from the dashboards. Spend time with your team — understanding where they are, where they could go, and what you need to build to get them there. The technology will continue to improve on its own. The people won't develop themselves.
