How Instacart Is Using Physical AI to Rebuild the In-Store Experience | NRF 2026

Hello, everyone.

Welcome back to day two of nrf.

This is Omnitok Retail.

I'm Anne Mazinga.

And I'm Chris Walton.

And we are coming to you from the Vuzion Booth, number 4921, where we'll be all day today and tomorrow.

So come stop by, say hello.

Check us out if you have the time and are here at nrf.

In the meantime, we're bringing you a lot of coverage of the people who are starring here at nrf.

Including.

Starring.

All right.

Including the man sitting next to me, David McIntosh, the chief connected Stores officer for Instacart.

David, welcome back to amitalk.

Thanks for having me.

It's great to.

Great to see you.

I think last time we chatted was Barcelona.

Barcelona.

Very different temperatures.

Summer, winter, but nonetheless, great.

Great to be here.

Yeah, a lot, lot different food, too.

Tapas, you know, very, very different heavy.

Meals at New York.

I prefer the pizza here, to be honest.

Big New York pizza guy.

Yeah.

All right.

Nice, nice.

All right, well, we'll start off like we always do, and I think it's good.

It's been a while since we've talked to you.

We have talked to you before, but it's been a while.

So let's give our audience a little refresher about who you are, your background, and also, what does a chief Connected Stores Officer actually oversee.

Yeah.

Happy to share more.

So at Instacart, Connected Store is really about taking what our customers love about the online experience, the convenience, the personalization into the store.

And a lot of our retailers came to us and said, hey, Instacart, you brought me online.

You powered my storefront Pro, my e commerce website.

You brought me online with fulfillment solutions.

I also want to digitize the store.

I want to figure out how do I drive increased loyalty signups, how do I generate more retail revenue, how do I drive a more personalized experience in the store?

And so Connected Store really sits at the intersection of those two insights and is a full suite of solutions designed to bring retailers online.

Smart carts called Caper.

Digitizing the deli with Foodstorm Loyalty Systems.

Carrot tags, which digitizes the shelf and actually integrates with Fusion so pickers in the store can more quickly get to items on the shelf.

Solutions like Store View that start to help retailers build an understanding of the shelf.

A whole set of solutions that really connects together to digitize the store.

That.

That.

That was a great overview.

Thank you for setting the table for us.

So my question is, you have all of those offerings that you're you're providing to your retail partners, what's the thing that retailers are still wanting or you that they're still dealing with, that they, they need your help with that you're.

Hearing most often I think this year one of the big things that we've heard is really AI, right?

And that's very top down in that what's happening is even to take CEO of a retailer, their board is telling them hey, I need a real AI strategy.

I want to see a real AI strategy next quarter, the quarter after that.

And so at a strategic level, retailers are really focused from our perspective now figuring out what's their strategy, what are they going to do this year, what are they going to do next year and what are they going to do for the next five plus years.

Right.

And so that's why you might have seen late last year, maybe fallish, we announced something called AI Solutions where the idea is to provide AI technologies for retailers that help them automate all aspects of their store.

There's agentic capabilities so they can actually embed agentec commerce experiences in their websites in their apps.

There's the ability to distribute those retailers to the hyperscalers like the Geminis of the world, right.

Eyes there's in store technology like Store View which enables our half a million shoppers to scan the shelves so the Ricohs can understand what's out of stock.

It also taps the side facing cameras of, of capers.

Right.

Get it up to the minute.

Every 20 minutes these cards go by the shelf, understand what's, what's on the shelf.

And then finally agentic analytics to make it easier for retailers to take all this data they're getting about their store, the sales, what's going on on the shelf and actually act on it.

So that's really great from a retailer's perspective then because you kind of are the one platform that they can use to kind of have that connected view with AI supplementing the data that you're getting from that store in a way that retailers can just plug and play like any of the other services you're offering.

It's like we can just add Instacart, have them take care of this for us.

That's absolutely right.

One of the things that we've been seeing in the last year is really the growth, gravity and the seamlessness developed between the solutions.

And what I mean by that is a retailer might for example deploy Storefront Pro, which is an e commerce solution and then they'll say hey, I now want to digitize my deli so that people can order ahead from the app and then they'll say, hey, I've digitized online and I've digitized the deli.

Now I want to digitize the store with Kaper or vice versa.

We've seen the opposite where a retailer might start with Kaper and say, hey, I now want to unify this with my online experience.

I get Storefront Pro.

And so we're seeing the same thing play out with AI, where for AI to succeed, it has to be deeply integrated into the data sources, into the operations of a retailer.

And a lot of retailers don't want a collection of 10, 20, 30 different point solutions.

They want something that all works together.

And I think the other thing I've observed really in the last year is you've started to see physical AI get more on the rate of retailers.

And you've seen Jensen talk a lot about this concept of physical AI.

And so what is physical AI?

Well, unlike AI in the cloud, where it's very, very large language models running on massive GPU clusters, physical AI is AI running in the store.

And so it's a very different technical approach.

So if you look at the Kaper card as an example, we have an Nvidia Jetson board and every single one of those carts and that Jetson board is running the algorithms in real time that understand what's going in the cartoon, what's going out of the cart, what's happening with the side facing cameras, what's happening with the location cameras, how do you make recommendations on the screen in real time to customers where the WI fi is often spotty, the WI fi often doesn't work.

And so you need those algorithms to run at the edge.

And I think what we're starting to see is that retailers realize that five to ten years from now, for them to avoid having the LLM hyperscalers like the Geminis of the world disintermediate the relationship with customers.

They need to build the best understanding of their customers in the store.

Right.

Because if you think about it, Gemini OpenAI anthropic, they're not going to put hardware like paper into stores.

They're not going to have.

Good point, right.

That understand where customers are in the store.

They're not going to have the clickstream on the caper carts of what people are doing, what they like, what they don't like, where they dwell, what they don't dwell.

And so all of that physical AI data that's then fueling this physical AI ecosystem, we're starting to see retailers really grok that that will be a key asset as part of that broader 5 to 10 year AI strategy that they're being asked by their boards to put together.

Got it.

You know, I gotta be honest with you, if I step back like I was, I was a little surprised when you said AI that they're coming to you and asking for AI.

I mean, I guess when I think about it, it makes sense.

But my question off of that for you is have you had to rejigger the trajectory of your connected store strategy at Instacart?

Or how, how have you had to reconcile what you were doing before?

Or is this just more of an extension of that?

It's definitely more of an extension.

Okay.

And I think if you look at for example, our decision to put Nvidia Jetson boards in these carts, okay, that was a decision we made years ago as we started to put the cart into mass production because we saw that vision playing out, we saw that AI was becoming increasingly more important.

Retailers would want to automate more of their stores and understand their customers.

And we really said we need to make this really deep investment so that we can best understand the store.

Our North Star is really building the best understanding of the users and the shelf and putting that together in a way that delivers a personalized experience for customers.

But then for retailers, they're getting maximal basket lift, they're getting maximal retail media revenue.

I think the other thing you're seeing, by the way, is that a lot of the AI technology in the last six to 12 months is accelerating what we're doing.

So if you think, for example, about the technologies around basket accuracy and anti theft within VA burner, we're able to apply techniques today with visual language models, for example, that just straight up didn't work six months ago.

And so we're really seeing a lot of this technology as a tailwind.

That's been a big theme of the show, actually.

AI enabling things that we couldn't do before.

So what I take from what you just said is it wasn't really a step change to your connected structure, is really an augmentation of your connected store strategy.

Is that a good summary of that?

Definitely an augmentation.

And I do think there's some areas that are step function.

Like take the cart assistant, for example.

Cart Assistant is a chatbot that will exist both in retailers, websites and on paper.

And it talks to each other.

So this cart assistant will learn about you, your preferences, what you like, what you don't like.

It'll build meal plans for you build it on a budget, you can use that in the retailer's app and then you can actually sync that to the Caper cart.

And so then not only will Caper start to remind you as you go to the store and say, hey, this item is actually not part of your health plan, you have an allergy to this thing or just proactively as you're going to check out, it'll say, hey, you forgot to buy the Tide detergent.

Like, you know, you normally buy it every month that's not, you know, in your cart and vice versa.

It will take everything you're doing in store.

What are you putting in your cart?

What are you taking out of your cart?

Where are you dwelling?

What are you not dwelling?

And feed that overall large language model to best personalize improve the online experience.

So that's an example, that's a step function capability that we just couldn't offer before that is accelerating both our e commerce business and the capers of the business.

So David, how are you seeing that connected stores are really moving the needle?

The technology in stores are moving the needle when it comes to operational efficiency.

Like what's the outcome?

Let's talk about some of the successes that you've seen from the retailers that are on Instacart, certainly.

And to step back to give you a little bit of an update since the last time we talked with Caper specifically as an example, we're seeing Kapir live now in 100 cities.

It's tripled year over year.

We have more than a dozen retailer banners using the technology.

A banner being like a Kroger, a Sprouts, a Wakefern, a Kohl's in Australia.

And this acceleration has really been driven by a couple things.

One is retailers have been increasingly measuring the basket left from Caper.

So we actually announced this on a prior earnings call.

Retailers are seeing double digit basket lift from Caper.

What they're doing is they're measuring, they're doing their own measurements.

They're looking at loyalty before and after.

What are you spending?

Running experiments and so forth.

So they validated customers are spending more.

Why are they spending more?

Yeah, they're spending more because of one, the running total, being able to actually track how much they're spending.

So many people underspend because they're worried about going to checkout, having to put things back.

But second, more importantly, it's the recommendations.

Right.

And on the recommendation side, what's exciting is now that we're scaling this product and have critical scale, we're Able to take so many of the techniques from the online side of our business to drive more personalization and basket lift for retailers and bring that on the caper side.

So at any given time we may have dozens of experiments running that ultimately maximizes personalization for users and drives more basket lifts.

So we're able to use all the sensor fusion data that we're getting to drive better outcomes for retailers on that side.

And what about the customer experience?

Like imagine this is now personalizing the customer experience and giving them more opportunity to engage with the capercart, upload my recipe and get walked through the entire store.

What are you seeing from like NPS data?

What are you hearing from your clients customers?

What are their thoughts on about the Instacart enabled or connected store experience?

Yeah, I think I mentioned last time one of my favorite things to do is store walks.

So you know, here on Saturday, after I mentioned the retailer demo, I just went to a couple different Wakefront stores.

You know, and with Wakefront, by the way, we're live in nearly 20% of all their stores with Wakefront here on the east Coast.

So, you know, I love going to stores.

I don't tell anybody I'm from instacart.

Right.

I'm just gonna.

Always a good plan, always a good plan.

And I'll even, I'll even prompt for negative feedback just to understand like what's actually happening in these stores.

And so first of all, I'll tell you that we're seeing double digit in store penetration, meaning a double digit percentage of the GMV is going through these carts and stores.

You can imagine multiplying that against the value.

When I talk to consumers, what do they say they love, first of all, the running total, being able to keep track of their spend.

They love bagging as they go.

They'll take the reusable bags, they'll set them up in the cartoon and you can take these carts directly to the car.

Right.

So a lot of people talk about automating checkout as skip the line.

And that's actually not what I'm hearing from consumers.

Yeah, that's what we remember last time we talked to you.

What they're saying is it's less work.

They're saying I don't have to take 50 items out of my cart and put it on the scanner or put it at the self checkout.

I can put it in the cart.

Once in, my bags go seamlessly my cart.

By the way, the carts are waterproof, snow proof.

So I saw this firsthand.

It was Raining, raining on the carts.

They're going through the ice.

So they're highly resilient.

So they can actually go outside the store.

People can use them like they do a traditional cart.

We've invested in stackable charging.

So the stores I visit have multiple fixed chargers.

The carts just stack like a normal stack of carts.

Oh wow.

Associates find that critical because if you don't do that, they've got to plug the carts individually, which then means associates don't like the product, they don't support the product.

And as you might imagine, associate support for the product is so critical, you need to make sure that the store is an advocate.

Which is why led by starting to talk about value levers.

Because the more value you're driving for retailers, the more they get behind the product.

I also heard from customers, they love the coupons, they love the recommendations.

We have caper exclusive coupons from cpgs that we're making available on the cartoon.

We sometimes will even activate with our retailer partners gamification.

So you can imagine a retailer might say, look, this is a customer that comes once every two weeks.

Can I get you to come back every week?

If you come back every week, I'll give you $2 off your next order as an example.

So customers love the gamification.

But I'll tell you the other thing that we've done to scale grocery is a business of edge cases.

In the last year, there's thousands of edge cases that we've ground down to make the product better and we've measured the impact of each one.

And so if you look at the dramatic improvements we've made in the product in the last year, it's not just the big rocks like better recommendations, more couponing, more gamification, more value features.

It's also just did you grind down the thousand little things to make it work?

Amazing.

And did you do that in a retailer specific way?

Well, the charging is a great example of that because that's a literal blocker to full scale rollout.

If you don't figure that out because you can't be charging all these cards individually.

So kudos to you for that.

I'm curious too because we're bringing this into the realm of the store.

Are the retailers still focused on their AI implementations with you through the caper cart or is that too simplistic of a question in terms of how to think about it for our audience?

No, I understand that it's a great question.

I would say retailers are taking a more holistic view in AI solutions similar to Connected Store.

It's a modular platform so a retailer can say, hey, I just want to do card assistant with Instacart, or I just want to do agentic analytics.

But all the solutions connect to each other and feed to each other.

And they are enhanced with physical AI investments like kpr, because then the car can understand the customer better, the store better, and make the whole system better.

So we do have retailers, like we announced, for example, when AI Solutions went out, that both Kroger and Sprouts were going to use card assistant.

And in their own announcements, they talked about how they expected to to use that both online and in store with kpr.

So we're definitely seeing retailers adopt the whole suite, but we're also seeing some retailers pick and choose individual technologies and start there before taking the whole stack.

Well, David, let's close you out with kind of what you're anticipating this year as we head into 2026.

As you continue to grow the business, scale the business, what would you be most happy to tell us next year year at NRF 2027 that you accomplished this year?

Great question.

I think number one is global.

I mean, that's part of global global expansion.

You know, you saw us launch on Australia with Kohl's.

We recently announced that we're coming to the UK Morrisons with Kaper.

We're here at NRF really talking to retailers about our entire connected store and e commerce products.

And we're seeing a lot of international retailers recognize our leadership in North America and they want access to our technology.

So to be clear, we have no plans for a marketplace internationally.

We're leading with our enterprise technology, our Storefront Pro, our Caper cards, our Food Storm technology, the AI solutions that we've developed.

So a year from now, you'll continue to see us really accelerate global expansion.

Number one.

That's great.

I mean, if you want to Accomplish More in 2026, go ahead.

I think, you know, the second thing I'll say is that we really are going to continue to enhance the value levers of Caper.

You'll see us with basketlift continuing to improve the recommendation algorithms, the ranking algorithms, and we're in such an exciting spot where you can now leverage this physical AI technology in a way where you can improve the personalization experience.

And as you mentioned earlier, often the technology is improving every month.

So I'm really excited as that is a tailwind behind us.

You'll continue to see us invest in retail media, more and more formats on Caper.

We announced at a prior earnings Call.

Engagement rates on Caper ads are similar to our online formats.

Right now is the engagement is sort of the foundational currency of an ads business.

You'll see us take the scale that we have online, $950 million in ads and other revenue the prior year, and really bring that scale into the store.

And then finally, you'll see us really push on, unifying the online and in store experience, really delivering that omnichannel experience.

What's so exciting about that cart assistant I mentioned earlier is we now have the technology to really deliver on that vision of you start your experience online, you continue in store, you go back online.

Our vision is that in five to 10 years, customers won't have to choose between shopping and store it online.

It'll be one single unified mode.

So three goals.

That's right.

Global expansion, increasing retailer value levers, and then really the omnichannel experience, unifying overachiever.

Here, look at this.

Don't ever.

I'm never comparing my 2026 goals list with you, ever, ever, ever.

David?

Yeah.

You should see my vision statement every year.

I don't want to.

I publish a vision statement.

I am confident that you.

I'd love to be your manager, though.

God, yeah.

Deliver that every year.

All right.

All right.

Well, thank you so much, David, for taking the time with us.

Thank you again to Viewan for helping us bring you all of our coverage here from nrf.

We will be here the rest of the day at booth 4921, so come stop by.

And until our next interview, be careful out there.