Welcome to Smart Consulting Sourcing.
This is the podcast where we talk about how companies actually buy consulting — not how the process looks on paper, but how it really works once you add people, politics, vague objectives, and the occasional “we’ll figure it out later”.
I’m Hélène Laffitte, CEO of Consulting Quest, where we spend most of our time helping organisations make sense of their consulting spend — and of the projects behind it.
If you haven’t already, I’d recommend subscribing — this is the first episode of a series, and it’s very much meant to be followed over time.
Today’s episode is called How to Buy Consulting in the Age of AI: Smarter, Faster, Still Human.
And I want to start by addressing a concern that seems to come up every time AI enters the conversation.
No, AI is not going to replace procurement.
At least not in consulting.
If anything, it’s finally going to make procurement sound like it has read its own RFP.
Consulting is not a category where you can automate your way to brilliance. Every project is different, the problems are rarely well defined at the start, and half of the value sits somewhere between judgment, framing, and knowing when a stakeholder is asking for clarity — or for cover.
So when people say AI is going to “disrupt” consulting procurement, they usually mean one of three things. Either they haven’t bought consulting recently, they’re confusing speed with thinking, or they are themselves consultants for AI implementation.
That said, AI does change something important. Not by taking over decisions, but by giving procurement more bandwidth, more structure, and — occasionally — the courage to ask better questions.
In this episode, I’ll explain why AI won’t replace buyers, but can be a very effective sparring partner. We’ll look at how it helps procurement challenge fuzzy briefs, stress-test assumptions, and make sense of complexity without pretending everything is suddenly objective and data-driven. And we’ll also talk about its blind spots — bias, hallucinations, and that impressive ability to sound confident while being completely wrong.
This series is not an ode to algorithms. It’s a practical, slightly cynical field guide for procurement leaders who want to use AI as a mirror, not a crutch. Because used well, AI doesn’t just save time — it forces discipline. And discipline, in consulting procurement, is already a small revolution.
So let’s start at the beginning. Why AI won’t replace buyers — and why that’s actually good news.
Why “AI Will Replace Buyers” Is Such a Persistent — and Lazy — Idea
Let’s start with this idea that refuses to die: AI is going to replace procurement.
More specifically, that it’s going to replace buyers when it comes to consulting.
It’s an appealing idea. Clean. Efficient. Comforting, in a way. If buying consulting were just about processing information faster, comparing options objectively, and selecting the “best” answer, then yes — AI would probably do a decent job. Possibly a better one than most humans, especially on a Monday morning.
The problem is that buying consulting has never worked like that.
Consulting projects don’t start with a clear problem statement. They start with a feeling. Something isn’t working. Something needs to change. Someone upstairs wants a transformation, but nobody has quite agreed on what that means yet. And somehow, procurement is expected to turn that into a scope, a sourcing process, and a rational decision.
That’s not a data problem. That’s a human one.
When people say AI will replace buyers, they usually imagine procurement as a purely transactional function — something that takes inputs, applies rules, and produces outputs. That version of procurement does exist. It’s just not the one that deals with consulting in any meaningful way.
In consulting, the hardest part isn’t analysing proposals. It’s framing the question in the first place. It’s understanding what the business is really asking for, as opposed to what it’s comfortable saying out loud. It’s spotting when a project is about solving a problem, and when it’s about buying reassurance, legitimacy, or time.
AI is very good at working with what’s written.
It’s much less good at dealing with what’s implied.
It doesn’t hear the hesitation in a steering committee meeting. It doesn’t notice when three stakeholders agree enthusiastically — but for three completely different reasons. And it doesn’t feel the political weight behind a “we’ve always worked with this firm”.
That’s why the idea that AI will simply replace buyers in consulting procurement is not just wrong — it’s a misunderstanding of where the value of the buyer actually sits.
The buyer’s value isn’t in pushing the process forward. It’s in slowing it down at the right moment. Asking the uncomfortable question. Pointing out that the scope doesn’t quite match the ambition. Or that the ambition doesn’t quite match reality.
And this is where things get interesting.
Because while AI won’t replace that role, it does change the conditions around it. It removes a lot of the busywork. It structures information faster. It forces clarity where there used to be vagueness. And sometimes, it asks the question that procurement didn’t dare to ask — because now it’s “just the tool”.
So the threat isn’t that AI will replace buyers.
The real threat is that it will expose which buyers were only ever managing the process — not the thinking.
And that’s exactly where we’ll go next: how AI, used properly, doesn’t automate procurement — it sharpens it.
AI as a Sparring Partner: How It Forces Better Questions (Whether You Like It or Not)
If AI doesn’t replace buyers, the obvious next question is: so what does it actually do?
And the honest answer is: it changes the dynamics of the conversation.
Not by being smarter than procurement, but by being relentlessly annoying in a very specific way.
AI has no problem asking basic questions over and over again.
What’s the objective?
What does success look like?
What’s the constraint?
What happens if nothing changes?
Humans tend to stop asking those questions much earlier — usually because someone senior has already answered them once, or because asking again feels… inconvenient.
This is where AI becomes useful. Not because it knows the answer, but because it doesn’t get tired of the question.
Take something very simple, like the early stages of a consulting project. The brief usually starts as a collection of slides, emails, and half-formed ideas. Everyone agrees the project is important. Fewer people agree on why. Even fewer agree on what would count as a good outcome.
When procurement uses AI at that stage, something interesting happens. You can take all that input — the messy version — and ask the tool to structure it, rephrase it, or summarize it back to you. And suddenly the gaps become visible.
Not because AI is insightful, but because it’s literal.
It forces the question: Is this actually a scope, or just a list of intentions?
Is this a problem statement, or a strategy wish list?
Are we buying expertise, capacity, legitimacy — or all three without saying it?
And sometimes, it does something even more basic — and more useful.
It helps procurement understand what is actually being said.
Not because stakeholders are unclear on purpose, but because business language has a talent for sounding precise while remaining deeply ambiguous. You get objectives that are actually symptoms, constraints disguised as preferences, and requests for “support” that are really requests for arbitration.
Running that input through AI — asking it to rephrase, simplify, or explain the brief as if it were meant for someone outside the room — often reveals a quiet truth: half of the confusion isn’t strategic, it’s linguistic.
AI is very good at translating corporate dialect into plain logic. And once you see that translation, it becomes much easier to ask the right follow-up question — or to realise that three people have been asking for three different things under the same heading.
That’s uncomfortable. And that’s exactly the point.
AI is very good at holding up a mirror. And procurement doesn’t always love what it sees.
The same thing happens when challenging stakeholder input. Traditionally, pushing back on a brief can feel political. You don’t want to sound obstructive. You don’t want to be “that person”. But when you use AI to test the logic of a brief — to ask whether the objectives are coherent, whether the assumptions hold, whether the timeline makes sense — the challenge becomes impersonal.
It’s no longer procurement questioning the business.
It’s the tool highlighting inconsistencies.
That small shift matters a lot.
It gives procurement cover.
It gives procurement structure.
And occasionally, it gives procurement the confidence to say, “If we launch this as is, we’re going to get exactly the kind of proposals we’ll complain about later.”
AI also changes how procurement prepares itself. Buyers no longer need to spend hours formatting, rewording, or consolidating documents just to get to a usable starting point. That time can be spent where it actually adds value — thinking through trade-offs, identifying risks, and deciding what really needs to be decided.
And this is where the idea of AI as a sparring partner becomes useful.
A sparring partner doesn’t fight the match for you.
But it forces you to sharpen your arguments, tighten your logic, and notice where you’re weak.
AI does that by asking naive questions with alarming confidence. And while that confidence should never be trusted blindly, the questions themselves are often exactly the ones procurement should be asking anyway.
The danger, of course, is when people mistake structure for judgment.
Because AI will happily generate a beautifully written scope that makes absolutely no sense in reality. It will summarize flawed assumptions with the same enthusiasm as solid ones. And it will never tell you, “This is politically impossible,” or “This stakeholder will never buy into this.”
That’s not a bug. That’s the boundary.
So used well, AI doesn’t make procurement faster by skipping thinking. It makes procurement faster by forcing thinking — earlier, more explicitly, and sometimes more honestly than the organization is comfortable with.
And that’s why AI is not a replacement for buyers, but a provocation. It pushes procurement into the role it always claimed to have: not just running the process, but shaping the decision.
Which brings us to the next step in the journey — the place where AI is most tempting, most useful, and most dangerous at the same time: writing the RFP.
From Blank Page to Sharp Scope: How AI Helps You Write a Real RFP (Not a Frankenstein One)
Let’s talk about RFPs — because this is usually where good intentions go to die.
Most consulting RFPs don’t start as documents. They start as compromises. A slide from strategy, a paragraph from finance, a few bullet points from IT, and a deadline that assumes clarity will magically appear on its own. Procurement’s job is then to turn all of that into something coherent, while being told — often in the same sentence — to be fast, flexible, and very precise.
This is where AI can genuinely help. Not by writing the RFP for you, but by forcing you to confront what you’ve actually been given.
If you feed AI a pile of stakeholder input and ask it to structure it, something useful happens. The tool doesn’t know what matters politically, but it is very good at exposing logical tension. It will happily point out that your objectives don’t align with your scope, that your timeline contradicts your ambition, or that you’re asking for both innovation and zero risk in the same breath.
That’s not AI being smart. That’s AI being literal.
Used properly, it becomes a very efficient way for procurement to challenge the brief before it becomes public. Not by saying “this doesn’t work,” but by saying “if we publish this as is, here is what it actually says.”
And sometimes, the result is even more uncomfortable: the realization that there isn’t really a scope yet. Just a direction, a hope, and a vague sense that consultants will help figure it out.
Which is fine — as long as you’re honest about it.
AI is particularly good at helping procurement clean up language without sanitizing the problem. It can rephrase, simplify, or restate a scope in plain terms, which makes it much easier to spot what’s missing. Success criteria that aren’t defined. Deliverables that are aspirational rather than measurable. Constraints that are implied but never stated.
This is also where AI helps procurement stop playing editor and start playing architect.
Instead of spending time rewriting sentences, buyers can use AI to test different framings. What happens if we describe the problem differently? What if we separate diagnostic from delivery? What if we admit uncertainty instead of pretending it doesn’t exist?
You can do this quickly, quietly, and before anyone outside the organization sees the document. Which means procurement can come back to stakeholders not with opinions, but with options.
And this is the key difference.
AI doesn’t turn a weak brief into a strong one. But it makes weaknesses visible early enough to fix them. And that alone changes the quality of what comes back from the market.
Because here’s the uncomfortable truth: most “bad” consulting proposals are not the fault of consultants. They’re a very rational response to a confused RFP.
If the brief is fuzzy, consultants will fill in the gaps — with assumptions, frameworks, and a healthy layer of marketing. If the objectives are contradictory, proposals will be too. And if success isn’t defined, everything will look successful on paper.
AI won’t stop that by itself. But it gives procurement a way to intervene before the damage is done.
Of course, there’s a trap here — and it’s an easy one to fall into.
AI will happily generate a beautifully structured RFP that looks solid, sounds confident, and is completely disconnected from organizational reality. It won’t tell you that the stakeholders aren’t aligned, that the timeline is politically driven, or that half the scope depends on decisions that haven’t been made yet.
That’s why AI is a drafting tool, not a truth serum.
Used well, it helps procurement move from a blank page to a sharp scope faster, with fewer blind spots and fewer late surprises. Used badly, it just produces a more elegant Frankenstein — stitched together, impressive looking, and still very much alive.
And once that RFP is out in the world, the next temptation appears almost immediately: using AI to find the “right” consulting firms.
That’s where things get really interesting.
Searching Smarter: Using AI to Expand (Not Shrink) Your Supplier Universe
Once the RFP is ready, the next temptation is obvious.
You have a clean scope, a clear structure, and now a tool that can “find the best consulting firms” in seconds.
This is where AI starts to look dangerously impressive.
Ask it for firms in strategy, transformation, data, sustainability — whatever your topic is — and it will give you a list. Polished. Confident. Reassuringly familiar. The usual suspects, plus a few names you half-recognise and feel good about not recognising too much.
And this is exactly where procurement needs to slow down.
Because what AI is really doing here is not mapping the consulting market.
It’s mapping the visible consulting market.
AI is very good at what’s already been written, indexed, optimised, and repeated. Which means it naturally favours firms that are large, loud, well-marketed, and very good at describing themselves as “leading”.
That doesn’t make the list wrong.
It just makes it incomplete.
The risk is subtle but real. If procurement relies too heavily on AI at this stage, it doesn’t expand the supplier universe — it standardises it. You end up with a shortlist that looks modern, data-driven, and slightly more diverse than last year… but still very safe.
And safety, in consulting, is not the same thing as relevance.
This is where human judgment has to come back in. Not to dismiss what AI produces, but to interrogate it. Why these firms? What kind of work do they actually deliver? Who will really show up on the project? And just as importantly — who isn’t on the list, and why?
AI can help procurement surface names faster. It can help cluster firms by capability, geography, or stated expertise. It can even help identify patterns — for example, which firms tend to show up together in similar searches.
What it can’t do is separate marketing fluency from delivery depth.
It doesn’t know which boutique quietly does exceptional work because they don’t spend time feeding algorithms. It doesn’t know which partner consistently over-promises, or which firm sends the “A-team” to pitch and the “B-team” to deliver. And it certainly doesn’t understand chemistry, trust, or political fit.
That’s not a failure of AI. That’s a reminder of what procurement is actually there to do.
Used well, AI becomes a starting point for supplier exploration, not a verdict. It gives procurement a broader map, but not the compass. The compass is still experience, networks, and the ability to ask questions that aren’t written anywhere.
And this is where the real opportunity sits.
AI allows procurement to say, “Here’s the obvious market.”
Human judgment allows procurement to say, “And here’s what’s missing.”
When you combine the two, you don’t just get faster sourcing. You get smarter sourcing. You widen the conversation instead of narrowing it, and you avoid mistaking algorithmic confidence for market truth.
Which brings us naturally to the next challenge — once proposals start coming back. Because sorting, comparing, and analysing them with AI looks very tempting… and slightly dangerous.
Evaluating Proposals with AI: From Bias Buster to Confident Nonsense Detector
Once the proposals start coming in, AI suddenly looks like a miracle.
Ten documents. Sometimes twenty. Different formats, different structures, different ways of saying roughly the same thing. This is where procurement has traditionally spent a lot of time doing work that is necessary, but not especially enlightening.
So yes — this is one of the areas where AI genuinely shines.
It can summarize proposals quickly.
It can compare scopes, pricing models, timelines, staffing assumptions.
It can highlight differences that are hard to see when you’re ten documents deep and slightly regretting your life choices.
Used properly, AI becomes a very effective bias buster.
It doesn’t care about brand prestige.
It doesn’t get impressed by glossy language.
It doesn’t assume that the most expensive option must also be the most strategic.
At least, not by itself.
The problem starts when people forget what AI is actually doing at this stage. It’s not evaluating proposals. It’s organising information. And those two things are often confused — sometimes deliberately.
AI will happily tell you which proposal is the most comprehensive, the most detailed, or the most aligned with the stated scope. What it cannot tell you is whether that scope made sense in the first place, or whether the proposal that looks “weaker” on paper is actually the one that understood the problem best.
And then there’s confidence.
AI is extremely good at summarising nonsense with total composure.
If a proposal is vague, contradictory, or built on heroic assumptions, AI won’t raise an eyebrow. It will rephrase it politely, balance it against the others, and move on. Which means that without human judgment, AI doesn’t eliminate bad thinking — it just standardises it.
This is where procurement has to stay firmly in the driver’s seat.
Used well, AI helps buyers prepare better conversations. It allows procurement to go back to stakeholders with structured comparisons, clear trade-offs, and very precise questions. Why is this timeline half as long? Why is this team twice the size? Why does everyone assume the same dependency is “out of scope”?
That’s valuable. That’s strategic. And that’s exactly where procurement should be.
Used badly, AI becomes a decision shield. A way of saying “the tool ranked it this way” instead of owning the choice. And that’s a very tempting shortcut — especially when the decision is politically sensitive.
But consulting decisions are never neutral. Pretending otherwise doesn’t make them more objective, it just makes them less accountable.
AI can help procurement remove noise, reduce bias, and surface inconsistencies. What it cannot do is decide which risk is acceptable, which compromise is worth making, or which team will actually work with the business once the contract is signed.
That’s still a human call.
So the smart way to use AI here is not to ask, “Which proposal is best?” but “What should we be debating?” And that’s a very different — and much more interesting — question.
Because once you get that right, AI stops being a judge and becomes what it’s actually good at: a very efficient assistant helping procurement turn information into insight, and insight into a real decision.
Which leaves us with one last topic before we wrap up this manifesto: where AI stops being helpful altogether — and where procurement absolutely needs to stay skeptical.
Where AI Stops — and Why Procurement Still Matters
At this point, it should be fairly clear where AI helps — and just as importantly, where it doesn’t.
AI is very good at structure.
It’s very good at speed.
It’s very good at forcing clarity where there was vagueness.
What it’s not good at is judgment.
It doesn’t understand when a project is technically sound but politically fragile. It doesn’t know when a stakeholder is aligned in principle but not in practice. And it has no instinct for when a “reasonable” compromise today will quietly undermine delivery six months from now.
Those things don’t live in the data. They live in experience.
This is where procurement still matters — and where it actually has an advantage that’s been underestimated for years. Procurement sits at the intersection of ambition and constraint. It sees patterns across projects, remembers what went wrong last time, and understands how decisions play out beyond the steering committee deck.
AI doesn’t replace that. It sharpens it.
It gives procurement the ability to say, with more confidence and less friction:
“Let’s slow down here.”
“This assumption doesn’t hold.”
“We’re about to optimize the wrong thing.”
And perhaps most importantly, it removes a convenient excuse. You can no longer hide behind complexity, or volume, or lack of time. When AI takes care of the mechanics, what’s left is thinking. And that’s where procurement either steps up — or gets exposed.
So no, AI won’t replace procurement. But it will make it harder to pretend that procurement is only about running a process.
Which, frankly, is long overdue.
Wrap-Up – Closing Episode 1 & Setting the Series
Consulting Quest 
