Dr. Elouise Epstein on Why Procurement's AI-Native Platform Era Starts Now

‍

In this episode of Procurement Unplugged, Fabian Heinrich, CEO and Co-Founder of Mercanis, speaks with procurement futurist Dr. Elouise Epstein, Partner at Kearney, about the next era of procurement technology.From paper-based sourcing and Excel-driven processes to AI agents and multi-agent systems, they explore how procurement has evolved over the past 25 years and why the industry is now entering a new AI-native platform era.They also discuss why legacy procurement suites still dominate many organizations, why point solutions may become obsolete, and what CPOs must do today to prepare their teams for an AI-driven future.Key topics:

• The four generations of procurement technology
• Why legacy suites still dominate procurement
• AI agents and the rise of autonomous procurement
• Platform vs. point solutions
• What CPOs must do to stay competitive in the AI era

‍

Fabian Heinrich (00:00)‍

Hello Eloise, a very warm welcome. It's a pleasure to have you here at Mercasphere. It has already been a very exciting morning with you, hearing everything about the future of procurement. Let’s sit back for a moment, zoom out a bit, and maybe discuss a few other topics.

‍

Dr. Elouise Epstein (00:19)

‍Well, first of all, thank you for having me here. I've really enjoyed being here. And yeah, let's talk about it.

‍

Fabian Heinrich (00:26)

‍How did you first come into procurement? And what did procurement look like back then?

‍

Dr. Elouise Epstein (00:32)

‍Well, I'm old, so I came into procurement the way many people in my generation did. We didn’t actually set out to work in procurement.

I originally came from a technology background and ended up joining a startup, a reverse-auction startup called Abbreviate.It was founded by people from Kearney and was eventually acquired by Kearney.

So that’s how I fell into procurement. From there, I worked my way up into consulting, and I’ve now been with the firm for almost 26 years.

‍

Fabian Heinrich (01:14)

‍That’s an exciting story, going from a startup with an entrepreneurial spiritall the way to becoming a partner at Kearney. But I imagine the way procurement worked back then was extremely different.

‍

Dr. Elouise Epstein (01:25)

‍Yes, very different.

I entered procurement around the year 2000. That was just after strategic sourcing had really come of age during the 1990s. Back then, you would have a team of maybe 25 people working mostly in Excel and before that even on paper, literally filling out paper surveys.

Just to be clear, that was the 1990s.

And those 25 people might be sourcing something like office supplies. Today we don’t even source office supplies that way anymore. But at the time you had big consulting teams doing exactly that.

So when I joined Abbreviate, the first thing we tried to do was digitize all of that, electronic RFPs and reverse auctions.

I happened to enter procurement right at that moment when things were moving from paper and manual processes toward digitization. That transition took quite a while.

As I mentioned earlier today, procurement technology first evolved through fragmented point solutions. Then those solutions consolidated through acquisitions, that was the second generation.The third generation was the explosion of startups.

And now we find ourselves on the edge of the fourth generation.

So in many ways, I’ve seen the whole evolution and I love it.

‍

Fabian Heinrich (03:00)

‍That’s very interesting.

Let’s dig a bit deeper. On the one hand, we’ve seen an incredible number of technology solutions emerging over the past 15years.

But on the other hand and maybe this is just my frustration as a technology entrepreneur, I sometimes feel like we are still where you described 20 years ago. Many things are still done in Excel.

And in procurement it sometimes feels like the same suite providers from 20 years ago still dominate the market. Some of them probably still have something like 80 percent market share.

‍

Dr. Elouise Epstein (03:41)

‍The old providers, especially the big suites, and I don’t even need to name them you know who you are.

They operate largely on inertia and fear. Many of them maintain strong relationships with CIOs or sometimes even CEOs, and that keeps the same technologies in place.

But those technologies never really solved the problem.

If you’re forced to use a legacy suite provider, sometimes you’re honestly better off just using Excel.

But that world has to change.

‍

Fabian Heinrich (04:20)

‍Then the next era was what many called the best-of-breed era. You were also a big ambassador of that movement and so was I.

The first company I built, Scalpia, was very much part of that best-of-breed wave.

Looking back now, why do you think best-of-breed was not the solution, or at least not the final solution?

‍

Dr. Elouise Epstein (04:44)

‍To be fair to myself, I always said that we ultimately needed a platform.

The platform was supposed to be the center of the spider chart, with point solutions around it. But nobody actually built thatplatform during the 3.0 generation.

Now people are starting to build it. But in an AI world, the way you build that platform is different.

Now that we have AI, you no longer need all those individual point solutions. That’s exactly what agentic systems are disrupting.

You don’t need provider X, provider Y, or provider Z anymore.

‍

Fabian Heinrich (05:18)

‍So what you’re saying is that many of the point solutions from the best-of-breed era are becoming redundant with the arrival of AI agents.

‍

Dr. Elouise Epstein (05:29)

‍Exactly.

And the problem is that many of those companies are financially underwater. They raised huge amounts of capital when capital was cheap, but their revenues never grew fast enough.

Now they’ll never be able to generate the returns investors expected.

Just think about all the companies on those procurement technology maps, how many raised Series A, Series B, maybe 10, 20,30, 50, even 60 million dollars.

But their revenues are nowhere near where they would need to be.

‍

Fabian Heinrich (05:59)

‍But would that actually concern CPOs? Whether a vendor’s equity is underwater or not?

‍

Dr. Elouise Epstein (06:05)

‍It should.

That’s a basic financial risk consideration.

And that’s exactly why we need to move toward real platforms.

‍

Fabian Heinrich (06:16)

‍I completely share that platform vision. At Mercanis, that’s exactly what we aim to build, one of those few true 4.0 platforms.

But recently, when I talk to CPOs, some say something interesting. They tell me: “Twenty years ago someone pitched the same platform vision to me.”

How do you respond to that kind of frustration?

‍

Dr. Elouise Epstein (06:45)

‍First we have to acknowledge that people have good reasons to be skeptical.

I actually wrote an entire book about the failure of the big suites to deliver on their promises.

And it’s not only the tech vendors’ fault.Systems integrators played a role. Business consultants played a role. Even Kearney, I’ve criticized Kearney for this as well.

We built bad business cases.

The analysts also played a role, many of them get paid by the vendors. Everyone had incentives to maintain the status quo.

Systems integrators pushed companies toward the big suites. Analysts reinforced that narrative.

But nobody really focused on solving the underlying problem.

So first we have to acknowledge that this model didn’t work.

Then we have to prove that something new actually does work.

That responsibility lies with people like me — analysts and influencers — and with entrepreneurs like you building new platforms.

We have to prove it before we sell it.

‍

Fabian Heinrich (07:54)

‍So basically: walking the talk.

The promise that wasn’t delivered for 20 or 30 years, now you’re saying that with new technologies we can finally build the platform that people hoped for decades ago.

‍

Dr. Elouise Epstein (08:09)

‍Exactly.

But again, nobody should take my word for it. And nobody should take your word for it either.

We have to demonstrate it.

When I work with organizations, I try to prove that anyone can learn to build an AI agent or adopt new technology.

And you’re building a platform that aims to finally deliver on those promises.

But we have to make it real.

‍

Fabian Heinrich (08:43)

‍I think people are an important point here.

Technology alone can only get you so far. If Excel is still one of the most widely used tools in procurement in 2026 and Excel was first launched in 1989, then clearly people also play a role in why new technologies have not been adopted more broadly.

How do you see that?

‍

Dr. Elouise Epstein (09:15)

‍It’s multifaceted.

Yes, people have not always adapted quickly. But to be fair, IT organizations and so-called digital organizations have also failed.

They’ve been busy implementing big platforms, Microsoft, Salesforce, Workday, SAP, Oracle.

They install those systems and then say: “Done.We implemented the platform.”

But that’s not the world we live in.

So we either need to move IT organizations forward, or replace them entirely, to bring better tools into companies.

At the same time, we also need to bring people along and prove that this isn’t just another version of Procurement 2.0.

This time we are actually moving the needle.

‍

Fabian Heinrich (10:15)

‍I think over the past decades companies also lost trust from employees.

At home people use smartphones and tablets with incredibly modern software. Technology evolves very quickly in their private lives.

Then they go back to work in procurement and suddenly everything feels like the 1990s again.

‍

Dr. Elouise Epstein (10:31)

‍Exactly.

And what message does that send to employees?

It basically communicates: “You are not valued.”

Why should someone feel motivated in that environment?

And we also have to consider generational shifts.

‍

Fabian Heinrich (10:45)

‍That’s actually a very interesting thought, the value of employees.

In a way, companies remain attractive to talent only if they provide modern technology.

Otherwise people will simply leave.

‍

Dr. Elouise Epstein (10:58)

‍Exactly.

Imagine you’re 30 years old today. Would you really want to work at a company that forces you to use SAP or Oracle for the next five or ten years?

If you follow that path, your skills will slowly deteriorate.

Then when you apply for another job, the candidate who has been building AI agents every day will have a much stronger story than someone saying: “Well, I used Excel and SAP.”

‍

Fabian Heinrich (11:27)

‍So in a way the question becomes: are you even still employable if you onlywork with record-keeping tools like that?

‍

Dr. Elouise Epstein (11:32)

‍Exactly.

But I also want to highlight another point you made.

In our personal lives many people already use AI tools in ways that corporations either don’t understand or wouldn’t officially allow.

Very often the junior employees are using those tools every day, while the senior leadership doesn’t really understand the technology.

So they don’t even realize what’s happening.

‍

Fabian Heinrich (12:09)‍

The generational shift is real.

Dr. Elouise Epstein (12:12)

‍Yes.

The older generation often says: “We have a policy, don’t do this.”

But younger employees don’t always absorb or follow those policies strictly. They use tools like ChatGPT to become far more productive.

And leadership often doesn’t even check.

There were actually two public cases involving Deloitte, one in Canada and one in Australia, where government projects discovered that AI-generated reports had been submitted without proper review.

It became a public embarrassment.

But that wasn’t ChatGPT’s fault. It wasn’t even the junior employees’ fault.

It was a leadership failure.

Senior leaders delegated the work but didn’t provide oversight.

‍

Fabian Heinrich (13:21)‍

Very interesting.

Imagine I’m a CPO in 2026. What would you advise me?

There are so many moving parts — technology stacks, talent, organizational change.

And many CPOs probably don’t want to admit that the suite or best-of-breed solutions they introduced five or ten years ago might not be the right solution for the next decade.

So what would you advise them?

‍

Dr. Elouise Epstein (13:58)

‍One thing I always find interesting is what happens at events like DPW.

Look at the CPOs who attend. They are the ones building the future.

Often the people on stage are actually members oft heir teams, people like Sam from Mars, Adam from Mars, Avanya from Coke or CCEP, or Gid Niten from GSK.

But their CPOs are there as well.

What those leaders do is create an umbrella that says: “If you work here, we’re going to do exciting things.”

If you go to many other conferences, you don’t see that same connection between leadership and innovation.

‍

Fabian Heinrich (15:03)

‍So it really starts at the top.

Leaders empower their people to explore new technologies and build new solutions.

‍

Dr. Elouise Epstein (15:10)

‍Exactly.

If CPOs are not attending events like DPW or conferences like this one, how would they even know what’s happening in the market?

They wouldn’t.

That’s why I constantly encourage CPOs to go to DPW, to walk through the expo hall and see all the innovation.

‍

Fabian Heinrich (15:36)

‍Another debate I’ve heard recently is about entering the AI era.

Some companies ask whether they should keep their existing legacy suite provider and simply add an AI layer on top.

There are several AI-layer companies growing extremely fast.

So the question becomes: do I rip and replace everything with a true 4.0 platform, or do I keep my existing systems and add an AI layer on top?

How do you see that debate?

‍

Dr. Elouise Epstein (16:27)

‍Let me answer that with a question.

If you’re a CPO, why would you ever want to do another Ariba implementation or another Coupa implementation?

‍

Fabian Heinrich (16:37)

‍Well — what if it’s already implemented?

‍

Dr. Elouise Epstein (16:39)

‍If it’s already there, fine.

But if you have a choice, why would you join a company that forces you to implement another legacy system?

There are many situations where companies spinoff divisions. In those moments they actually have the chance to start fresh.

But what do they usually do? They implement Ariba again. Or Coupa again.

They have a unique moment to change and they don’t take it.

That’s crazy.

Now, if you already have a legacy provider in place, yes, you can add an AI platform on top.

That’s possible.

‍

Fabian Heinrich (17:29)

‍And nobody has to admit failure. You just put lipstick on the pig.

‍

Dr. Elouise Epstein (17:35)‍

Exactly.

And honestly, that CPO will probably be fine. They might stay three to five years and then move on.

But if you really want to build something sustainable, you have to start creating a path toward a true 4.0 platform.

Adding an AI layer on top might work in the short term.

But you won’t get the real benefits.

I often use the example that OpenAI’s finance department is only about 18 percent the size of comparable companies.

A “pig with lipstick” organization might still need 100 people.

A true 4.0 organization might need 25 and achieve far more.

That’s the real difference.

‍

Fabian Heinrich (18:41)

‍So it’s not a long-term solution.

Looking ahead, if we meet again in one year, how do you think technology will have evolved?

What would you hope to see?

‍

Dr. Elouise Epstein (19:01)

‍Right now we’re seeing a lot of hype. So the first step is separating hype from reality.

What I expect to see is AI platforms like OpenAI, Anthropic, and Gemini expanding further into the cloud space.

They may start displacing parts of the traditional cloud infrastructure.

Today the big cloud providers focus on compute and storage. But the center of gravity is shifting toward AI.

And I think we will start seeing AI employees.

Imagine you run a chemicals company and you can hire an AI employee with PhD-level expertise, essentially an AI agent that performs knowledge work.

Why wouldn’t you do that?

Within a year, I think we’ll at least see the first real examples of that.

‍

Fabian Heinrich (20:14)

‍We’re moving toward that very quickly.

If you think about it, one year ago we barely talked about multi-agent systems.

Now we’re already close to deploying them in major procurement organizations.

So your vision doesn’t seem unrealistic at all.

‍

Dr. Elouise Epstein (20:38)

‍But again, it has to be framed within a platform architecture.

If you only add an AI layer on top of an old system, you won’t unlock the full potential of multi-agent architectures.

That’s why I strongly believe the future lies in true 4.0 platforms.

Even if you temporarily put lipstick on the pig.

‍

Fabian Heinrich (21:10)

‍Those are great closing words.

Thank you so much for being here and for this fascinating conversation.

I’m really looking forward to meeting again next year at Mercasphere 2027 and seeing what 2026 has brought.

‍

Dr. Elouise Epstein (21:32)

‍Thank you for having me. I’ve really enjoyed being here.

And these are exactly the kinds of events I love participating in.

So thank you for the invitation.

‍