Physical AI: What It Is & How IoT Will Combine With AI To Make Retail Operations Self-Aware | Spotlight Series

Foreign.

This Retail Technology Spotlight series podcast is brought to you by the Omnitalk Retail Podcast Network.

Ranked in the top 10% of all podcasts globally and currently ranked in the top 100 of all business podcasts on Apple Podcasts.

The Omnitalk Retail Podcast network is the network that we hope makes you feel a little smarter, but most importantly, a little happier each week too.

And, and this podcast is just one of the many great podcasts you can find from us here at Omnitalk Retail alongside our Retail Daily minute, which brings you a curated selection of the most important retail headlines every morning and our signature podcast, the Retail Fast Five, that breaks down each week.

The top five headlines making waves in the world of omnichannel retailing and comes your way every Wednesday afternoon.

Hello, everyone, we are your co host for today's interview.

I'm Anne Mazinga.

And I'm Chris Walton.

And Chris, this is a little bit like the teaser we have of Stranger Things coming back.

You know where you, you watch the first season and then you have the second season teased at the end.

Well, we are doing that today with part two of our series with Wilyat.

Now, we left you last episode with the cliffhanger on physical AI and today we're going to dive into that with Wiliot's director of data product and AI Doran Hassan Doran.

Thank you so much for joining us and welcome to omnitalk.

Thank you so much.

Thanks for having me.

You said my name really, really well.

So I, I'm practicing my Hebrew.

It's not getting any better, but I'm, I feel like if I just keep practicing, perhaps it will.

But we've got a lot of opportunity today.

Yes.

Practicing one name at a time.

All right, well, before we get started, before we dive into everything, let's have you.

Why don't you give maybe the.

Maybe there were some listening that weren't here for the first episode.

They weren't here for the cliffhanger.

We why don't you give our audience a brief reminder overview of what wiliot does and what your role is there.

So for Anyone new to WilIot, we are a pioneer of a technology called ambient IoT.

Think of it as a way to connecting trillions of everyday items.

You can think about products, cases, pallets.

So connecting that into the cloud with tiny battery free Bluetooth sensors that we call IoT pixels.

So these pixels, they continuously capture data about where the things, you know, where they are, what conditions they're on, how they're moving, how long they've been Idle and they do this automatically.

So no scans, no batteries, no touch points and all that data, the ambient data from IoT pixels, but also in addition to complementary data from systems like, you know, rfid, all of that data flow into the Wiliot intelligence platform where AI and ML transform that data into intelligence so you can think about real time inventory accuracy, freshness in size, shrink detection, dwell time analysis and etc.

Many, many, many others.

My role of the director of Data product and AI.

So I lead that physical AI strategy.

So essentially how we fuse that IoT data ML gen into living intelligence layer so retailers can trust run operations smarter, leaner, more predictably.

Doran, how do you explain that in in from Wiliot's perspective, like dive a little bit more into what physical AI looks like inside of a retailer that you might be partnered with or in concept even.

I guess just one note before we dive in into the real use cases is that I think I spoke about this a little before in our, in a blog post, but physical AI, it doesn't replace humans, it empowers them.

So I guess the ultimate model is a partnership model between physical AI and humans.

So you have ambient IOT required, right?

Exactly.

Humans are required so we can think about this.

The system is a decision support tool rather than an autonomous decision maker because humans, you know, we have the, the business context that regardless like AI and, and, and ambient won't have.

So it's a combination of these three.

And now to your question about real use cases.

You probably can demonstrate it.

So I think the first thing that I, the first use case that I mentioned is, you know, inventory accuracy.

You can be more accurate with the, with where your things are.

So it resolves inventory discrepancies before they become stuck out.

So one of the retailers biggest pain point, it's also being used to automate restocking and replenishment.

So when something arrives, sits too long, gets misplaced, the system will notify you instantly.

So not in a report tomorrow, not during an audit next week.

It's in the moment.

So it turns replenishment into a guided, almost autonomous workflow.

That's the second use case.

The third one is, well the most straightforward one probably is freshness, cold chain integrity.

So ambient IoT by design monitors the temperature, humidity and we always add more capabilities here in William like light and other capabilities.

So we monitor that across trucks, the backrooms, sales floors, all this and the AI will come in and flag excursions early.

It would say hey, this pallet is warming up.

Or this case has been out of range, let's say, for 27 minutes.

So it's actually preventing the spoilage before a customer even sees it.

So you've mentioned a few terms already.

You've mentioned IoT, which is a term that, quite frankly, we haven't talked that much about on our show.

And in recent memory, you've mentioned IoT, you've mentioned ambient IoT as well, and then you've mentioned physical and digital AI, so.

So piece apart all of those for us.

Like, what is.

What is different about each of them?

And where does one begin?

And where do the lines start to blur as well?

I think as.

As Nick said before, Internet of Things, I think, is a concept that we're familiar with quite a while ago, right, this pretty much like 15 years ago, right?

Yeah, yeah, right.

Quite a while ago.

It's actually connecting the things, physical things, to the Internet, pretty much.

And it started many different forms.

And I think, as Nick said it correctly, what WILLIA does before we jump into ambientarity is focusing on that team, the things themselves, not just the Internet, of course, cloud and AI.

Now connecting these things, essentially letting them speak.

And ambient IoT is also not a thing that Wiliot invented, of course.

It's pretty much Internet of things and everywhere.

This is the stuff that we enable with the Bluetooth, because Bluetooth is pretty much everywhere.

So the Internet of things is.

You have your things connected anywhere, everywhere in real time.

So you've got IoT, which is basically like the system of connected devices, ambient IoT.

What I'm hearing from you is that's.

That's the kind of always on nature of the IoT, right?

It's ambient, it's always there, it's in the background, it's collecting data.

Right, exactly.

So this digital AI is what we know is pretty much what we're familiar with.

You know, AI is also a very old concept.

Now.

People talk about AI, you know, recently because of advancements, you know, recent investments like ChatGPT, but even before that, you know, advancements in computer vision and, and deep learning and models like this.

But AI is also an old concept and it's all based on digital stuff.

You know, it's.

It's, you know, it's being trained on the Internet on things, documents, texts, spreadsheets, numbers, all that kind of stuff.

But it's not really connected into the physical world.

Think about it.

Think about being trained on.

On physical things, on where the things are and their condition.

This is sort of the realm of physical AI.

So physical AI is like a.

It's like a derivative of the digital AI.

It's like taking digital AI, adding ambient IoT, connecting them together, and then you get that physical AI.

I love that.

I love how you said that.

It's like, it's like connecting the whole thing.

So, so basically, so what you're saying then is now we have the capabilities as retailers, or actually as people in general walking around in our daily lives to connect all these things together.

We can use the, the information we're gathering from the ambient IoT sources, combine it with the data we're getting from either digital processing of information, say large language models, for example, or even computer vision, AI, camera systems inside of a store, tracking and identifying people or products in space.

We can combine that all together to create this new idea that you, you guys are calling physical AI.

And the manifestation of that.

Absolutely.

That leads into my next question is, you know, what is the consumer seeing?

What is the shopper seeing as a result of a retailer who's made this investment in these, these multiple points of contact into the digital AI, the physical AI and the Bluetooth beacons and Internet of things that they have working here in the store?

What happens for me when I, when I'm in the, the store, I think.

The simplest way to look at it is, is this everything retailers are, are doing with physical AI, it ultimately shows up as just a better shopping experience.

So technology may probably sit around behind the scenes, but the benefits, they land squarely with you, with the customer.

I think the magic of physical AI is that improves customer experience without you even being knowing like you don't knowing what physical AI even is.

I can give you some examples so we can dive in.

So first, as you as a shopper, you'll see fewer out of stocks and more consistent availability.

You know, so when the retailers, they know exactly what's in the building, what needs to be replenished so the shelves to stay full.

So you and me and I suffer this a lot.

You will waste, waste less time hunting for items.

You know, take me, takes me a lot of time.

Always when I go to places, I, I say, I wish that I could just speak to AI to find that thing that I'm looking for.

You know, it's just, yeah.

Second, you know, you'll get more fresh food, you know, because I mentioned, you know, the, the freshness and, and, and that status of food.

Of course, monitor temperature so you'll get better produce.

You know, the entire process of click and collect that home delivery experience, it will be better because, you know, when retailers have the real time visibility, they can assemble orders more accurately so they can ensure that items show up together and in the right condition.

You know, it will mean also fewer errors for customers overall, fewer misspeaks, mystery delays and all this.

And again, it will be invisible to you.

You're not going to see sensors, you're not going to see any data pipelines.

What you'll notice is just a better experience.

I mean, I think this is really important too.

Now Doran, as we're talking about the, the number of households that are looking for especially fresh foods inside of grocery right now.

I think retailers have to be even more diligent about making sure that they have good inventory, they're reducing waste and that they're getting as many products on the shelves because you know, you guys just had a big launch with Walmart recently and that's something that we're hearing from consumers in the news.

Like they are going to Walmart because they're able to get the things that I've been out of stock on other shelves, other grocer shelves because of things like not having this full visibility to the location, the temperature, the status, you know, the, the placement of all the products within their store.

I imagine that's something that you are hearing too as you are rolling out to more of these Walmart stores.

Is that true?

Absolutely.

Both from the retailer's perspective and from the customers, shoppers.

Course this whole conversation reminds me of that old adage that, you know, good omnichannel retailing.

When you don't see it, that's essentially what you're saying to Roan and, and, but I'm, but the funny thing about that is like from a customer perspective, that's true and that's what we want.

We don't want the customers to see it, but on the flip side, the people that we do want to see it and understand the efficiencies are the store operators.

So, so tell me, how does this make the day to day lives of the store operators or the store employees better?

I said it before, but physical AI doesn't replace people, it empowers them.

So let's think about warehouse and store teams, store managers.

The biggest impact similar, I think to the customers is simplicity.

And I think the most straightforward thing to think about is just less manual work.

So for them, for these teams, it's no more searching for missing pallets, walking around scanning things or guessing what to restock, manually checking temperatures.

The system does all of this automatically.

We can also think about, you know, you can think about a Live operational picture.

So it gives the operators a sort of clear real time picture of what's happening in the warehouse of the store.

So for example, the associates will know what just arrived, needs to be moved, what's out of range, what's sitting for too long, of course, freshness, compliance to get better.

So they don't need to scan and only to see if something is out of compliance because they would know.

It's very hard to know.

Sometimes we have just really great examples.

Sometimes you would see strawberries, for example.

Strawberries, you can see them well and they go, let's say for example, they go into the truck, they being mishandled, so they're being frozen accidentally.

They're not supposed to be frozen, but something in the AC of the truck doesn't go well.

And when they leave the truck, they're not being frozen anymore.

And this actually, this process of being frozen and not damages the quality of the strawberries.

So if you don't have that tracking, you'll be in the store and you see, oh, strawberries look good, but in fact they're not.

And with real time, with physical AI, you can track that temperature in real time and the AI will detect it and tell you, hey, this is not fresh or you should not use it.

It's cool because it gives me this vision of like something that I've always wanted to see like in retail, like a huge mission control board.

Like in those movies where you see like people that operate train lines or subway lines where they see all their trains going in all these different directions that they know where they are at all times.

It's the same way we could think about inventory.

Like we can understand all the exceptions that need to be managed throughout the day in a given store at a given time.

And that's just, this is really cool to think about.

So, all right, so I'm curious then.

So like both, all that said, so if we get into like, if we buy into the idea of physical AI, like how should I as a retailer prioritize my investment to make this kind of mission control vision possible?

Like where do I start?

Where do I go next?

As I'm thinking about investing in the various solutions that bring this about your.

Comment about that, you know, big screen that you can actually see in that store where everything goes, I genuinely believe where going there.

And yeah, this, you know, it's, it's a long path.

We get there and, and to answer your question, we start with a small step.

We're not going to jump ahead into something that is is too complex because I think this is probably a mistake that retailers would do because we have an amazing technology.

Ambientality is tremendous and amazing and gets better.

And of course AI gets better, the models get better, so their combination is better.

And it's very tempting into jumping into many use cases and trying different things and trying to scale them fast.

And this is where I think we should do differently.

I think the, you know, if a retailer is willing to invest in this, in physical AI, we should start small, should start identifying a real problem with a real business value, scope it well and use physical AI in order to solve that problem.

If you solve that problem and you can see it generates some ROI and value, then you can scale into different use cases.

I think that's the way, the right way to invest in this.

We've been talking to a lot of retailers lately and one of the things that keeps coming up, especially in the grocery space is there's a lot of point solutions right now.

There's a lot of dashboards like you were just talking about that are kind of feeding this main brain for the store operations of a store.

As the, as the head of product here for what and what you're developing and building with Wiliot's technology, as you're thinking about bringing in this, this concept of physical AI into retailers, how do you think about how the information that you're gathering fits into or you know, is the, is a puzzle piece within the greater ecosystem for the store's operations?

What are you doing in your work to make sure that it can be easily integrated?

Well, I think everything we develop at product and the data products that we develop, we try to connect that into a real use case or real problem.

I think Nick mentioned this before, we learned a lot from deployments like Walmart.

So those learnings, they're built directly into the product.

So simplifying infrastructure requirements or faster onboarding or integrating directly into the customers systems, you know, that's one, one thing we think about, I think Nick mentioned, you know, Walmart's native systems that we integrate into.

So we keep that in mind all the time.

The other point, the other point I want to get your take on here too because I think this has become, it's been confusing for me and I imagine is for some of our listeners too is there's the idea of like we say starts, we say start small.

Right.

You know, we say that but at the same time also we want to get, we're also leery of like just creating a set of just point to point solutions.

That are just all over the place and you don't know where you're going.

So the real thing here is to start small with a platform approach.

Right.

That you can then build upon that platform with more and more use cases.

Is that the right way to think about this?

Absolutely.

Sometimes the hard things, it's like we can think about it like in physics you have static friction and, and kinetic, I guess, friction.

The static friction usually the initializations and things.

Integrations, security checks, integration into systems.

This is what the Wiliot intelligence platform actually solves.

It gives you a really accessible and easy way to manage your assets, manage your solutions, manage your insights.

And you start there and you start with a small solution.

Then you know, finding a new solution and scaling.

This becomes easy when you start with something very dedicated.

That's a really key point.

I'm glad I.

Okay, thank you for that.

Yeah.

Doran, how do you think about what's next in your roadmap as you're thinking about product and data and AI converging?

What's on the roadmap for you and for Wiliot as you head into 2026?

Super exciting stuff, to be honest.

Yeah, I bet.

I think we're entering a moment where, you know, physical, the physical AI world intelligence becomes the new competitive edge.

Not just digital analytics, not just, you know, forecasting or all that, but real time sensory, ground, truth, data.

I just think, you know, that, you know, retailers who adopt this now are just building automated workflows like us, you know, smarter supply chains, almost self aware operations.

I can't share too much yet, but what.

But we're unveiling what we're unveiling.

I guess we're heading into 2026, which huge momentum and one of the biggest milestones is NRF this January 2026.

So we'll be announcing major updates.

Both are IoT pixel technology, those little pixels computers, I guess and will it intelligence platform can't share too much yet, even though I really want to.

What we're unveiling represents a significant improvements, you know, in performance, in range, accuracy, AI processing, both attacks and the platform itself.

And just you know, as important importantly, we're making the platform easy, easy to deploy, easy to integrate, easy to scale across different large distributed retail networks.

Right.

And with all those Walmart stores, all of the Walmart stores that you're deployed in right now, like you said earlier, I think you're going to be gaining a lot more information about how to continue to build that physical AI infrastructure at an enterprise retailer and then figure out how to do that in a variety of other retailers as well.

We are so thankful that you had the time to spend with us today.

You mentioned that you'll be out at nrf, so if you don't mind sharing with our audience the best way that they can get in touch with you either before or at NRF this year.

So, first of all, if you're attending nrf, be sure to visit us.

We're going to be at booth 3469 and we'll be giving live demos.

We'll showcase new capabilities.

All the stuff that we mentioned, you know, it's exciting stuff.

So it's definitely demos, real stuff.

You'll see.

You'll see, I guess, physical AI in action, that's going to be in RF.

If you want to reach me, LinkedIn works under my name or my email, my work email.

The Ron hazan.com I can't wait to see.

I can't wait to see what you guys are going to unveil.

I'm dying to see what this intelligence kind of platform idea is because then, you know, it gets to the point you said, like, build one use case and then sprout out from there.

Which is the right approach, folks, if you're listening, like, that's how you got to approach this opportunity of physical AI and how it's going to impact retail.

So.

All right, well, that wraps us up.

Thanks to Duron of Wiliot for educating us today on this new concept of physical AI.

We got into it quite a bit there and thanks to everyone, as always, for listening in.

Today's podcast was produced, of course, with the help of Ella Seward.

And as always, on behalf of all of us here at omnitalk, as always, be careful out there.