Accenture on AI, GLP-1s, and the Future of Grocery Retail | FMI 2026

Hello, welcome back to FMI.

This is OmniTalk Retail.

I'm Anne Mazenga.

And I'm Chris Walton.

And we're coming to you live from the simbi booth, number 118 here in the FMI tech section.

I want to give a big shout out to Simbi first off for helping us bring you all of our interviews from the show here at fmi and our next guest that we have joining us back on the show for FMI number two, we have Karen Fang Grant Kyle.

Karen is the managing director and Global Research lead at Accenture.

Karen, welcome back.

Thanks for joining us again.

Thank you for having me.

Yeah, we're excited to talk to you.

Yeah.

Karen, when I reached out to you, I remember saying to you, I was like, you were one of my favorite conversations from last year.

I feel like I learned so much from you in terms of how you approach the business and the insights for the industry.

So I was kind of thinking maybe it'd be good to start off with like having you talk through some of the trends you're seeing or that we're seeing and have you read and react to them too.

So the first one top of mind for me is GLP1s.

Like how big is that?

What are you seeing?

How are you seeing consumers respond to it?

The whole nine yards, wherever you want to go with that.

Oh, it's gigantic.

It's one of the biggest.

Gigantic.

Really?

Yes.

Well, you think about the fact that GLP1s are an injectable, a very expensive injectable, not easy for consumers to access, and yet it has had so much impact that in the US we've seen obesity rates drop from about 40% to 37%, which is not enough, by the way, but a huge change in a very short period of time.

Right.

That's like almost 10% drop.

That's great.

Yeah.

So what is the impact on that for the grocery industry then you think as you look forward or for the consumer in terms of changing tastes or behaviors?

Well, consumers themselves are.

Then, you know, those ones who are using them are seeking to find those macronutrients, let alone the micronutrients they need because they're eating less and so they're struggling to get things like protein, which is why we see protein in everything.

So you saw recently, Starbucks had released the cold protein foam.

That's just yet another example of how companies are addressing those consumers needs for macronutrients.

And you see it on both new and existing products.

So you see a re skinning of old products to say to highlight those nutrients that consumers are really looking for.

Yeah, we're seeing it even here in some of the sample rooms.

They've got protein all over them like protein cookie dough, I noticed cookie dough ice cream.

Protein cookie dough ice cream.

I love the, you know, give us all the things we want.

Both protein and that delicious indulgence that we need.

So that, so you said we went from 40 to 37.

Are you expecting that to accelerate?

Like, I mean, I know the, the, I don't know what the right word is, but the mechanism to take the drugs are still changing too.

Are you expecting that growth curve to continue to accelerate, to impact the industry even more?

I think it will only accelerate.

I mean I think there, it's such a, there's huge pent up demand for this to the point where the drug for which, you know, the use case for which it was originally designed, which is for diabetics, there are a lot of diabetics who still struggle to get access to that drug.

And so I can only see supply increasing.

And that's not just in that injectable form as you mentioned, but also there's a pill form that has just been going through reviews and makes it even easier for consumers to be able to take this.

Well, Karen, one other trend or buzzword that we've been hearing a lot about, and I'm curious to get your perspective briefly is on agentic commerce, how you think about that and how you think that retailers and brands should be thinking about agentic commerce.

But then outside of that, what other trends in tech are you.

We'll cover that quickly.

But then what other trends are you hearing about in tech that we should be paying attention to?

I love how you say we'll cover agentic commerce quickly because it is the topic of the hour.

I was just at NRF last week.

It was just blanketing the entire convention center and every conversation, all the discussions that Google had kicked off with their uct.

So many Google announcements, everyone's really excited.

And that's driven by the consumer.

So the consumer, I mean, it's funny because I was just in a session where they're saying, do we trust AI or not?

Broadly, we don't in personal senses.

We do in our own research where 36% of AI users consider AI to be a good friend.

This is what they say.

And this is not.

They're naming it.

This is not techno geeks.

This is not weirdos like us.

These are people who use AI as little as once a week.

They're saying it is a Good friend.

And that's because this is a technology that is more human than any that we've faced.

So everyone has to respond because the LLMs are seeking to move into the space of commerce.

That's where the money's at.

They have to recoup some of that great investment that they made in their models.

And everyone is sort of figuring out then what is the kind of natural, logical ecosystem that creates value both for the players and for the end consumer who plays what role and who controls what part of the design you're making.

So what advice would you have then for retailers who are thinking about how to take the next step in their agentic commerce strategy?

Yeah, it's a combination both of get in early so you learn it because things are changing so quickly.

It is a different mindset instead of, I don't know, paradigms than you might have thought about your industry in the past.

But don't try to be everything.

I think a good example of this is Walmart where they are experiencing, experimenting very mindfully in different areas.

They have Sparky, their own agent or shopping assistant, but they are also working with OpenAI and they're basically laying out their cards to see what roles different players will play because I don't think those players themselves know what they're going to be doing.

So they're testing across the board.

And then what about other technology outside of agent E commerce?

Let's talk about some of the things that are kind of bubbling to the side top for you as you think about, you know, what you'd be tracking or following or maybe doing more experimentation with.

I think we need to distinguish between that specific case of agentic commerce and broad AI.

Sure.

Right, right.

And I think this is something that people kind of, they've been.

A year ago or two years ago it was all gen AI and then it got brought into agents and now it's agentic commerce.

But we have not even scratched the surface of the possibilities with AI, including how it impacts the work we do every single day as brands and retailers and how we need to fundamentally rewrite all of those processes and workflows in order to capture those capabilities.

Is there one area that you would apply AI to right now?

Like is it operations, is it planning, forecasting?

Like where inventory analysis, marketing?

Like where would you be telling people to kind of identify some of those main categories to start applying some the of of this AI technology too?

Yeah, yeah.

I mean so if you talk about planning and forecasting, AI has been there for a long time.

It's just how you apply it needs to change.

In the past, it was better predictions in the future or today.

And in the future it needs to be better recommendations, better next, best action.

So the planner pulls up their laptop or their computer and doesn't have to sift through all of the steps.

Moving the repetition.

Yeah.

And so basically the AI agent is giving them three decisions you need to make today.

Go.

Right, right.

That makes sense.

So, Karen, I'm curious too, because you're the global head of industry research.

Industry research for Accenture too.

Like, I'm gonna go back to agenta commerce before we move off it too quickly.

So when you're, when you're assessing the landscape here, how quickly do you think consumers are going to move to a willingness to have an agent purchase products for them?

Like, are we talking like, this is going to be really fast, like iPhone adoption rates, or is it going to be like a slower grow?

Like, and is it going to be different by geography too?

I'm curious what you think on that.

Oh, absolutely.

It's going to happen faster than I think.

Wow.

I don't know if that's a contradiction in terms or not, but let me nuance that a bit.

A lot depends on what the players in the field do and are able to enable right now.

It's a bunch of cobbled together little solutions that are actually not a great shopping experience.

It's not easy for consumers.

They aren't, you know, we aren't going to do it.

Speaking as a lazy consumer myself.

Right, right.

But they want to.

But they want to.

So consumers.

We did research in the spring where we asked consumers that exact question.

Would you let an agent that knows you and your preferences just go all the way to click to purchase and the product shows up at your doorstep?

Right then it was about 12% said, yeah, we would actually.

12%.

And this was a global audience that we looked at.

Okay.

When we asked in the fall, that number had risen to 15%.

Okay.

You know, a small incremental change.

But you think about, you know, the actual solutions that were in market that allowed a consumer to do that had not changed material.

Lots of exciting announcements, but nothing really realized up to that point.

And yet we see that number changing when it's still, I mean, we think about agentic commerce all the time.

How much does the average consumer.

They're just kind of.

They're still in a fairly experimental stage.

But when it happens and when those solutions exist, I think it'll happen very quickly because we are all looking for that easy Solution.

Right.

It was all theoretical at this point in time that you're asking them what were some of the reservations that you identified in that research too about using agented commerce for our audience?

So we know, like, what are some of the hurdles that people are going to potentially need to get over?

This is no different than any other kind of commerce.

So we did a lot of research on social commerce and we started to pay attention to the growth it had, particularly in other markets than the US but including in the US and we found that the problem was, you know, is it easy?

That's why I keep going back to do those mechanisms exist and our company is providing that ecosystem of services that make it easy for a consumer to actually get what they want.

Wow.

So this is one of those areas where the consumer is actually out in front of the technology in terms of desire.

Even though the technology is two years.

Old, I would say.

Or not even two years old.

Agentix, like, not that.

What yet?

No, it's not.

It's not that old at all.

And I mean, what was possible last week is different than what is possible this week.

Wow.

But I would say the, the consumer is well ahead in terms of their demand.

The tech is actually quickly catching up to that and creating new demands for the consumer.

Retailers and brands are a little bit behind.

Right.

And the tech companies too.

So everyone has.

That's why I keep going back to ecosystem.

Everyone has to figure out their right role and the right solution.

Wow.

And I imagine that once you can start like the buying only one product at a time right now.

Oh, goodness.

Yes.

Which is a big component of that too, is not how people shop for a lot of things.

They're doing a basket, like in grocery, for example, that will change.

It's rare that we're just buying one thing.

So I imagine that things like that allowing basket building and things like that on top will start to change.

That would be a hurdle, maybe.

And linking my basket for one occasion, like my weekly shop with my baskets for other occasions, like the fact that I'm going on vacation next week, that changes my weekly shop, actually.

But do the different systems talk to one another now?

Right.

Wow, that's fast.

There's some pent up demand here.

That's what you're getting at.

Wow.

Wow.

All right, so let's shift gears a little bit away from technology.

One of the things that always gets talked about at this show is private brands.

Yes.

How important?

I mean, we hear a lot about budget constraints, you know, depending on who you talk to.

How important Are private brands now going forward for grocers and for the consumer?

I think it's, there's, there's a good, there's a good, the bad and the ugly.

We always hear about it label where I think that consumer receptivity of private brands went up and continues to go up.

It's probably at the highest it's ever been, right?

Yeah, exactly.

And that's partly based on the value that it provides an individual consumer, but it's also based on the fact that certain offerings have been particularly good.

I always call out the Kirkland Signature brand at Costco.

People love it, I certainly love it.

And it's a super quality brand.

I would go for that reason.

I think the kind of bad part of it is that grocery retailers have not necessarily kept up with that again, pent up demand that consumers or receptivity that consumers have.

Some grocery retailers have provided some very good and distinguished offerings.

But there's still a lot, if you look behind the hood, a lot of let's look at, let's wait for, for those trends to exist and then jump on them with something that's a copycat product.

And maybe they do a really good job at that.

But their mechanism for scanning, sensing what could be new trends and then the ability to create those new trends I don't see as well developed yet.

And that's where speaking with my CPG hat on, the CPG companies definitely have an advantage because that is what they do all the time.

And so they are seeking in a larger extent to create those new trends in that new demand space.

So would you say that's a white space then for those that want to capitalize on it?

I do, yeah.

That's interesting.

I was having that conversation with the lead merchant at a pretty big grocer this morning and he said exactly the same thing.

So that's validated.

Well, if you think about grocery in the U.S. it's not a very distinguished offering necessarily one from another.

No.

Right.

If you go into a grocery store, you still find many grocery stores that have not changed in decades.

And could you really tell without actually reading the sign which grocery store you were in?

There are some great standout examples in our market and then there are many that are not.

And then the trick is scale too.

Right.

You need scale for private label.

So there's always the catch 22 there in terms of how do you get the money and the resources to do it.

Right, right.

Everyone always points at Europe and says, oh well, we should do it like them.

I'm like they have a far more concentrated market in their different countries than we have in the United States.

So it's a little bit easier.

Yep, great point.

Well, Karen, let's close with looking ahead to FMI 2027.

We're going to have you back on the show.

What do you think we'll be talking about then?

Will we still be talking Magentic commerce?

Will we still be talking AI or is it, is it something new?

What is still out there that you think will evolve over the course of the next several months?

We do this research with the World Economic Forum on the AI transformation of industries.

And in this, we did this, you know, a lot of interviews with different companies in the industry and tried to measure their progress towards something more than individual experiments or use cases.

You know, are they engaging in functional reinvention, actually rewriting the entire ways that say merchandising top to bottom actually work all the way through to.

Are they rewriting the enterprise?

You know, the way that merchandising works with supply chain, works with store ops.

And right then, it was still very early days.

Everyone was very much about experimentation.

What I'm hearing is that again, those capabilities, like we're settled in that this is a real thing.

We don't need to experiment to find out whether it's real.

We may need to experiment to be able to learn.

But it's time to get to actual scale in what kind of results we can achieve with these investments.

So what I expect to see are not just larger scale initiatives, but ones that touch multiple parts of the organization, completely rewriting the workflows of a given process, reshaping the workforces and reshaping people into new roles.

I was going to ask that, like, how do you think culturally then businesses will look or shift?

Like, do you see we saw chief AI officer as a role that was being hired on.

Do you think that that will fall, you know, within each individual's role at the C suite level and on downward or how do you think?

Like, what will, what will our workforce look like at the major retailers and grocers?

I think that there are things that will change and things that won't.

So when you look at the frontline workforce, there will still be a large frontline workforce.

Stores still matter, you know, as much as stores.

We talk about technology and agentic commerce.

People still want to visit stores and that's where the heavy volume is.

So we need those in store associates to be there, be present and be understanding of who their customer and their shopper is.

But they will be enabled in a different way.

So they might be able to recognize me when I come up or at least have as much information as I do, information as I do from my AI friend.

They should be equipped with that same knowledge.

I think another change that will occur is amongst the leadership.

So as we think about the reshaping of the workforce and removing maybe some of those, you know, distributive or management type positions that are kind of, you can see AI coming in to enable people directly on the front line.

So you cut out some of the middle layer.

Those individuals start to take more strategic roles.

And that means that that top layer of leadership has to be far more highly skilled and far more strategic than in the past.

Wow.

So, Karen, if I wrap up what you just said there, if I try to summarize it.

So like the world I'm envisioning actually is where you have a lot of organizations that are process mapping how they get things done and then saying where can the AI fit into orchestrate versus just trying to randomly pilot different improvement sites, but really looking at the process end to end.

So I see like a lot of process maps in the future of a lot of leaders to try to understand how do we get more efficient and that efficiency can be based on a lot of things, but AI being the tool to do that.

Is that what you're saying?

It is what I am seeing and I think it's actually a suboptimal result.

So I don't know if either of you have ever heard of Eric Brynjolfsson.

He's a professor, he was at, I believe, mit and he's now at Stanford.

Focuses a lot on AI and productivity.

And he wrote these great papers about what was called the productivity paradox.

Why have we not seen productivity gains from these new technologies that are truly transformative?

And he tells the story of how when you looked at old manufacturing plants that were steam powered and you tried to move them to electric power generation, they did not see the efficiency gains at first because they had not changed the way that the work was being done.

Electric can be distributed versus central, so you could actually break up the manufacturing and get these productivity gains.

We're seeing the same thing in the work really of Free Tail, where, you know, the industry is exactly doing what you're describing, taking existing processes and trying to optimize them.

And, and that's, that doesn't actually capture the capabilities of what we're moving towards, which are these agents that completely rewrite the workflows into something new.

They find the most.

There isn't a process.

You have to redevelop the entire flow.

You have to remap the entire flow.

And recognize that the agents, because they themselves are quote, unquote thinking, are actually mapping their own process in the most optimal way each time.

Right.

Wow.

We're gonna have you back.

Oh, my God.

That was so good.

All right.

This is so great.

I love this.

Well, thank you so much for giving us your time, your insights today.

Thank you again to Simbi for helping us bring you all of our coverage from FMI today.

You can stop by and see us in booth 118.

We'll be here all day.

And until our next interview, be careful out there.