Moving Beyond AI Experimentation with BCG X | CGF 2026

Hello everyone, this is omnitalk Retail.

I'm Chris Walton and I am coming to you live once again from the Consumer Goods Forums Global Summit in Vienna, Austria.

And we are of course in the CGF Leadership Studio, which is sponsored of course by our friends at Vuzion.

Now joining me is Nico Debelfonds.

Nico is the Managing Director of bcgx, which I can't wait to learn more about for EMEA and South America.

So, Nico, thanks for joining us today.

Thank you very much for having me.

Yeah, I'm excited to talk to you.

So tell us about yourself, your background and your role at bcg.

So I've been at BCG for a long time.

Too long, let's not go there.

But I have been working in AI in one shape or form for the last 15 years.

Hired the first data scientist in Paris in 2013.

Really?

And we've been building up a team of forward deployed engineers, AI scientists, product managers, designers who actually do AI.

They don't talk about AI for our customers and that is called bcgx.

And I manage that team for our EMISA region, which is Europe, Middle east and South America.

How did you get started in that, like, you know, 15 years ago of being, you know, into AI is pretty.

It's a pretty long time.

So how did you get into that initially?

Well, we were actually trying to solve some complex problems that frankly did not fit into an Excel.

You know, things like, you know, churn prediction or pricing in retail, personalization in airlines.

And so we started to leverage data science as a tool for that.

And we quickly realized that this was an extremely powerful capability that we needed to internalize.

And so this is how it started with a few people that then grew into a global organization of now 3,000 people.

3,000, Wow.

So basically grew out of business need then?

Essentially.

Yeah, exactly.

All right, so CGF and bcg.

You also just released a new report on AI as well, AI in the consumer industry.

I'm curious what prompted that report and can you summarize some of the major findings for our audience back home?

Of course.

So essentially what prompted it is everyone's talking about AI, but there is a lack of clarity from especially the sea levels that are here at the CGF on what is actually real versus the marketing.

Everything that you say to analysts but that are not okay.

So we looked at, on the CPG side, consumer packaged goods, and on the retail side, what is the true state of play?

Where are people driving value?

How are the leaders moving?

And what we saw is actually that one it's still at a nascent stage.

In CPG especially, 75% of CPG companies are stuck in pilot mode.

Only 18% have some form of scaled.

Impact, especially in CPG.

Yes.

In retail it's a very different story.

Wow, okay.

Because in retail half of the retailers have actually moved ahead and have some at scale impact, half of them have nothing.

So it's also a two speed world that is quite interesting to see feast or famine.

Yeah.

And that impact ranges from, I would say in the CPG side mostly right now in everything related to pricing, revenue management, which is an obvious place to start.

In retail it's mostly in supply chain at the moment.

But both are going into new directions.

CPGs are investing into AI for innovation and marketing.

Retailers are investing quite a lot in AI for the merchandising function which is really the core engine of the organization.

So both are on the move.

Wow, that's so interesting.

I mean even as I'm talking to you, like I'm hearing different things from different people throughout all the different conferences.

My third week of conferences, retail conferences in Europe and you hear something different around the perceptions of what's going on in the industry depending on who you're talking to.

So let's dig in some more then.

So I want to know too, what do you think is the biggest pocket of value sitting out there, particularly when it comes to demand generation and the use of AI?

Yeah, so again, so I think there are two kinds.

One is what can you do to transform your own organization?

And I would say in the CPG business, the beating heart of a CPG business is the innovation and marketing workflows.

The ability to transform these with agents and AI solutions of any kind, to dramatically increase the speed to innovation, the time to market of new product, the quality of these products and of the way you market them to consumers is potentially a game changer.

On the retail side, that is the merchandising function.

The ability to source the right products, get the right assortment at the right price on shelves and you know, better negotiate with suppliers is going to be a game changer.

So that is on the internal side, on the consumer facing side, the biggest thing of course is agent E commerce.

And you know, how do you make sure that your brand is discoverable, visible from the right consumers on all of the AI platforms out there.

And then how do you integrate yourself into the emerging agent E commerce ecosystem which is a completely new channel for all of these brands to master and for retailers, which is potentially either a new gateway or a new Competitor.

Yeah.

And how do you think about that on the retailer side too?

Because I've had everyone on my show from you should be rushing into it and experimenting like crazy to.

You should pull back.

Wait, get, get generative AI set up on your own site first.

You know, get.

Understand that make sure you're showing up on the LLMs where important but you know, maybe slow down the role on the agentic side of things like how do you look at that?

What do you advise retailers to do?

So I think one should you be experimenting?

Yes.

Because if you're not staying close to the pace of innovation, you're going to lose grip.

Now should you make very strong commitments in this emerging ecosystem?

Not yet, because the protocols are not there yet.

Now that being said, creating your own capability is a no brainer because this will give you proprietary insight into the way consumers are behaving, what they are actually prompting, I mean data.

There is no data out there to understand what consumers are actually looking for, which kind of consumer, for which kind of occasion, what kind of products, etc.

And it will also give you a way to get your, you know, your brand into the ecosystem.

So if you have your own agent then at some point you can connect your own agent with the emerging ecosystem as opposed to just being a participant like anyone else.

Yeah, it's like practicing before the big game, so to speak, in a lot of ways.

So you know, to that point though we have seen a lot of pilots come and go.

You mentioned there's a lot of pilots going on in both the retailers and the CBGs already you have experience, 15 years of experience of building an AI function and quite honestly an organization inside BCG.

What are the secrets from moving from pilot into actual scalable uses of the technology?

Yeah, so I think when you talk about AI for your internal operations, I think there are two traps you don't want to fall into.

The first one is dispersion.

So you know, if you see a roadmap with 200 use cases, walk away.

You know, success comes from.

I have a few bets that I drive and to end where I try to, you know, really change how people work.

And the second is the incrementality, you know, trap trying to just improve slightly.

The way I operate today, this doesn't lead to significant improvements versus I take something and I really change how people operate with a much more AI first logic.

Which means by the way that the roles of people will change, the skills that I need will change, the entire construct of the organization will change and it will require a lot of managerial input, but this is what will drive sustainable change in the business.

So that's a really good nugget.

So what I hear from an organizational perspective then if I'm leading one of those teams is like, you probably should be pretty aggressive on your stretch goals in terms of what you're trying to accomplish through the application of AI.

Yeah, you need to stretch yourself way beyond what is feasible today because the technology evolves at exponential speed.

Organizations evolve at linear speed at best.

And so if you set a target based on what is achievable today, it will take you three years to get there.

And in three years, the technology will have moved multiple times.

Therefore, you need to take a stretch ambition that takes into account the exponential acceleration of the technology so that you put the right goals, but also you need to be focused because managing that change across 20 different parts of the organization is going to be difficult, if not impossible for the leadership.

Right, right.

Okay, so looking at.

So then where are you seeing success?

Let's start in CPG first.

Where are you seeing the CPGs particularly have success or those that are differentiating themselves on the AI front?

Where and how are they differentiating themselves?

So where is, you know, the core critical functions of the company?

Revenue management, marketing, and increasingly innovation.

Okay, okay.

So this is where they are driving most of the impact.

And then historically also on supply chain, how is really a matter of, you know, as I said, of managerial focus, meaningful investments.

We should not kid ourselves.

You know, we're not printing money here.

This is, you know, the typical roi.

That's what our study revealed in our own work as well, is somewhere between 3 and 4x.

Okay.

It's not a, you know, 200x that is not happening because the amount of work to be done on the tech side, laying out the foundations, building all of these, you know, solutions that really transform the business, but also on the people side, on the change management side is actually quite meaningful.

Right now a forex return on investment on your tech investment is pretty good, but, you know, it requires meaningful investment to get to a significant impact and a meaningful amount of managerial attention.

This needs to be the top priority of someone high in the organization, 100% for the next two years.

Right, right.

Yeah, I got to think at least the next two years.

Right.

So probably ongoing.

All right, so let's get you out here on this because the one thing we haven't talked about, agents doing the actual work of or inside of organizations.

If I'm a retailer CPG CEO, how should I be Thinking about and preparing my managerial staff for the advent of when they begin to manage agents to do the work.

So I think lots of people are extremely skeptical about what AI can do.

There were a lot of other promises under delivery.

So the best way to do this is through lighthouses.

Lighthouses, you know, you take some parts of, you know, your organization that are aligned with your top priorities that I mentioned and you create the, let's say, dark factory equivalent of that, of that team.

And you need to take teams that opt into that because you know it's a meaningful amount of change.

You're trying to create a mirror of that with a fully agentic loop and as little human involvement as possible.

That shows, number one, the art of the possible, number two, the kind of impact you can get.

Number three, also what it takes to scale this.

And that allows you to start creating the backbone, the backbone of the agency platform, the backbone of the data, and then the backbone of capabilities that you need to transform.

Once you have this, then you can start scaling it in the right domain.

So let's say I have one lighthouse in parts of marketing, then I can expand across the end to end marketing function or even beyond that.

But you prepare the organization by showing the art of the possible and then preparing the said, let's say, underlying capabilities and enablers you will need to scale.

Yeah, so I'm curious, so the lighthouse approach, never thought about this and never asked it.

And so I want to make sure I articulate it correctly.

So do you take a, do you envision like a tiger team approach too, where somebody does say you put the lighthouse in one position, the team goes and goes.

After you get the people that want it, you get a line in the organization, you go after you make the change, do you then take the same people and apply it to a different area?

Or do you think you, or do you think you do it separately?

How do you think about that?

So the best way for me is to separate this into two parts.

There is one team that creates the backbone and that is one team.

Okay, so that's just consistent across everything.

Exactly.

Okay.

And then you have teams organized by either functions or end to end processes, depending on the organization.

End to end processes is better, but can be sometimes more difficult given the structure of the organization that is quite function led.

But.

And then these teams, they will transform bit by bit the entire thing.

It's the same team that you know because then you connect the dots across an end to end workflow or an end to end function and you can drive meaningful pnl, what you should not do is essentially have a team that jumps from marketing to supply chain, then to finance.

And because that, you know, you need someone who goes down the experience curve in that domain and connects the dots.

Yeah, you need somebody that has that expertise, but you need the people that get the foundation of how AI works.

Exactly.

To then bring that knowledge to all those other teams.

But that is key, though, and you.

Need to create the platform and the connectivity from a data standpoint.

Right.

You know, because ultimately some.

Not all, but some of the data and the knowledge and the context that these agents will need will be replicable across your different processes.

And, yeah, there should be scale in that effort if you're doing it correctly versus having to start over every time.

Wow, man, this was great.

Thank you so much.

Thank you so much.

We're going to have to do this again sometime.

All right, so thank you very much, Nico, for joining us today.

Thank you to CGF and Vuzion for helping us bring you all of these great interviews with all, all these fabulous retail and CPG executives all three days of the summit.

And as always, everyone out there, be careful out there.