AWS Mid-Year Leadership Summit 2025 | Rajiv Chopra, AWS

Clips
News
More from AWS Mid-Year Leadership Summit 2025

Rajiv Chopra

VP, AWS Healthcare AI, Life Sciences and Just Walk Out

AWS

Just Walk Out continues the self-service shopping journey for AWS

The concept seems simple, yet the technology behind it is remarkably complex. It’s pay-as-you-go, self-service in the purest form; walk into a store, pick up an item and just walk out.AWS’ Rajiv Chopra speaks with theCUBE about self-service during the AWS Leadership Summit.The technology side has been supplied by Amazon Web Services Inc. using a platform called Just Walk Out that was developed and launched in Amazon Go stores in 2018. The technology has expanded beyond the retail store use case, as explained by Rajiv Chopra (pictured), vice president of Amazon Just Walk Out at AWS

play_circle_outline Overview of AWS' Just Walk Out technology and its significance in retail.

play_circle_outline Seamless Shopping: How Just Walk Out Technology Transforms Retail in Hospitals, Stadiums, Airports, and Universities

play_circle_outline Importance of evaluating logistics and operational needs for potential implementation of Just Walk Out.

play_circle_outline The role of computer vision technology in achieving high accuracy in transactions.

Info
Transcript

Rajiv Chopra, AWS

Rajiv Chopra

VP, AWS Healthcare AI, Life Sciences and Just Walk Out AWS

In this AWS Mid-Year Leadership Summit interview, Rajiv Chopra, VP of Amazon Just Walk Out, joins theCUBE’s John Furrier to unpack the evolution and impact of computer vision in retail. Chopra shares how AWS has transformed the breakthrough technology behind Amazon Go into a scalable, edge-powered solution for partners across stadiums, hospitals, universities and airports. With over 250 deployments outside of Amazon properties, Just Walk Out is redefining how consumers shop by enabling fast, frictionless experiences without checkout lines.

Chopra detai... Read more

explore Keep Exploring

What advancements have been made in the application of computer vision-based technologies over the last decade for various sectors? add

What are the use cases for the technology mentioned, and how is it implemented in different environments? add

What benefits does Just Walk Out technology provide for businesses, and how should they evaluate its relevance to their specific needs? add

What are the unique features and differentiators of the Just Walk Out computer vision technologies? add

bolt Powered by CUBE AI

Rajiv Chopra, AWS

search

>> Hello, I'm John Furrier here in Seattle at AWS' headquarters for theCUBE and AWS Halftime coverage. We're at the midpoint of the year and so much has happened, we need the updates. We've talked to Matt Garman, all the executives and the leaders and the experts to get all the news and to really understand what's going on in the market. Generative AI has changed the game, new methods of models and efficiencies coming, new uses patterns, new ways to do things. And this next segment is my favorite, Just Walk Out technology. We've got Rajiv Chopra, VP of Just Walk Out at AWS. This is the walkout technology. Rajiv, thanks for coming on theCUBE, I appreciate it.

Rajiv Chopra

>> Hey, John. Thanks for having me here.

>> So explain what Just Walk Out is, because Amazon has so many kind of sensational breakthroughs, from drones delivering packages to the Go store. Just Walk Out, maybe to me it's associated with the Go, but this is an opportunity for AWS. Just Walk Out technology is an effort, it's a division, it's product, it's a technology. Explain what Just Walk Out is.

Rajiv Chopra

>> Sure, sure. John, what we've done over the last almost decade of investing in computer vision-based technologies is taken the core computer vision-based technology from the Go store and externalized as a product in different forms that business and retailers in completely different segments like travel or hospitals or stadiums or universities and their cafeterias, they can use that same computer-vision powered technology to make it be super simple for shoppers to tap a card, walk in, shop, and just walk out.

>> I think image; image search, image video search, from the web 2.0 days. You can kind of see and get, but the horsepower wasn't there. But now with edge AI and all the supercomputing capability that's now available, computer vision is the killer app. It's a lot of data, but it's clearly set up for watching everything. So why not Just Walk Out and have it charged to my account? Where is it use cases? What is it being used for? What's explain some of the momentum and use cases you're seeing now, and where does it go? This is a commerce solution. This is digital meets real world, computer vision, commerce. Explain the use cases and where does it go?

Rajiv Chopra

>> Absolutely. I think we are just getting started. I think we've obviously shown the technology in our Amazon Go stores. We've proven that if you, as a fan, go to a stadium and you want to quickly get back to your seat to watch the game, you don't want to miss anything there, you walk into a store, you don't have to stand in long lines. You tap your card, pick up something, and get going. Those are proven verticals. But now, if you think about that same scenario in a hospital as a medical provider or as a loved one for a patient, you have the expectation of being able to get a snack maybe late in the night when you're waiting there. And so, this technology in many different formats allows for stores that are unattended, allows for extended duration without needing additional staff, allows you as a business to be able to manage large spikes in customer, in shopper traffic while also reducing theft and losses because you're tapping, there is an expectation that this automated checkout is just seamless. And the staff that's in the store can now be focused more on greeting your customers and on taking care of upselling them goods as opposed to being a cashier.

>> One of the things I'm thinking about when you're talking is that yes, first of all, we can all relate to the don't stand in line for a beer or a hot dog at a game, to walking in and getting some stuff at the pharmacy or walking in for some service at the hospital. It's fast checkout basically. It's just walking in, walking out.

Rajiv Chopra

>> Yep.

>> The thing that gets interesting is you mentioned automation. I can envision robotics also being involved, because there's also a huge robotics wave where unattended attendance, robotic attendance could manage dispensing a soda, a beer, chips, pharmacy. So with that Walk Out, you have identification, you have an ID, there's verification.

Rajiv Chopra

>> Yeah, yeah.

>> And service.

Rajiv Chopra

>> You have a great point, that the opportunities are endless to try and combine the simplicity of digital payments, digital authentication, identification of the customer with the ability to serve customers as seamlessly as possible. I think that for businesses, whether you're in food and beverage or you are in apparel or retail, you can then pivot the associates and the staff to really take care of the customer, as opposed to things that can be easily automated. And I think computer vision technology is at a point where our ability to bring high accuracy models to the edge with our custom edge computer devices, our cameras, they're giving the accuracy that makes sure that the customer experiences high quality is something I'm very excited about.

>> This is a retail solution, so if someone's evaluating this, what should they consider? Because some people are like, "Whoa, I'm skeptical" or "Whoa, this is too good to be true" or, "wow, I want it tomorrow." So I can see the attitudinal scale there. What is the evaluation sequence of events look like? Can you share what happens when people look at this, and how should business leaders look at this?

Rajiv Chopra

>> Sure, sure. I would say obviously our solution solves a couple of business problems really well. And so, businesses should see whether in their store as retailers, whether these are the most important problems that they have to solve. Number one is obviously, like we mentioned, Just Walk Out technology works really well in high throughput, crowded scenarios with lots of people moving in. We can still with high accuracy know who's coming, who's taking what, right? So if you have a problem or a scenario where there is a lot of traffic and you don't want to spike up staffing in order to support it with long queues, but you want a seamless solution, that's an obvious fit. Another one is if you want to extend the duration or you want to have unattended stores. It's hard if you have low traffic, so you don't have a traffic problem, but you really want to extend the stores and want to optimize your staffing, Just Walk Out is a great solution there.

>> Yeah.

Rajiv Chopra

>> Another one, there are other self-checkout solutions, but how Just Walk Out, with its ability to guarantee the payments, can also take care if you have theft or shrinkage. That's another area where Just Walk Out can really help.

>> So what are some of the innovations? What's under the hood? What's the secret sauce? Computer visions? Is it GenAi? Is it algorithms? Can you share the differentiation and what makes this unique?

Rajiv Chopra

>> Absolutely, absolutely. I think, as you know, we've been investing in the Just Walk Out computer vision technologies for almost a decade now. But the latest versions of our computer vision algorithms are no different than all of the other generative AI techniques, the deep learning techniques, transformer models that we are basing it on. And over the years we've drastically reduced the number of models to handle the kinds of unique scenarios that we've tested in our Go stores and tested in our Just Walk Out stores. You imagine not just an individual coming into the store but a whole family coming in. Kids are there grabbing things, putting it back, people coming in with all sorts of hats. The ability to handle all of these scenarios using vision foundational model at the high level of accuracy, that's super cool. In order to make this happen, we've also had to invest in doing a lot of this vision algorithms on the edge. So we have our own custom edge compute devices that are highly optimized for doing vision algorithms. We have our own networking gear to make sure that we don't have any drops and any breakage in the streams that are being processed. And then, obviously it's backed up with the cloud, and so that's the way that we've invested in it.

>> It's incredible just to listen to it. It's almost like, okay, my mind's blowing because the magnitude of data, the scenarios you mentioned, the computer vision is not a small data ingest. It's massive. And you're streaming data across the network or processing on the edge, so you've got to have everything teed up to get this to work. It's kind of like talking to Prasad about the data centers. It's not just a facility, it's not a building. It's like an engineered system, the whole thing. You have to build this thing end to end.

Rajiv Chopra

>> Spot on, spot on. I think that's another value that we offer to our business customers, is we own the hardware, the data, the processing. We'll take care of it if something is not working. We come in and replace it. So it's an end-to-end solution that we offer. And yes, you can imagine a typical store has 30, 40, 50, and a large store might have 100, 200 cameras. And being able to ingest that data, be able to track hundreds of potential customers, shoppers in a store across it, the amount of processing is quite impressive.

>> All right. So what's the problem? Because I would see everyone want this. Is there a cultural change management going on in the market? Are you seeing rapid growth? Is there market development requirement? I just don't know because I haven't really been following the retail side as much. We were at NRF this year where we covered mostly the GenAI piece from a different angle, but we're covering a lot of the retail commerce piece right now. Digital is the cloud basically and it's edge, so we look at retail as an edge thing. But I don't know, I just don't know the progress. What's the momentum like? Can you share some of the business stats and highlights?

Rajiv Chopra

>> Absolutely.

>> With some numbers?

Rajiv Chopra

>> We've doubled the number of stores that we've launched. We are now over 250 stores that are all-

>> They're all retail stores?

Rajiv Chopra

>> These are all in healthcare.

>> So diverse stores, not Amazon stores.

Rajiv Chopra

>> Not Amazon stores.

>> Okay.

Rajiv Chopra

>> These are not Amazon. These are our customers working...

>> They must love it.

Rajiv Chopra

>> They love it.

>> What's up for them? Savings? Better customer satisfaction? What are the drivers there?

Rajiv Chopra

>> I think in the business-to-business-to-consumer part, we both always, our customers, our business customers have the end customer always top of mind. So that shopper experience and delighting, it's a magical experience. That is what is the first attraction for our customers. Obviously, the ROI they get from being able to increase revenue. A lot of these stores, for example, if I look at Hudson News in the airport, right? They have a traveler coming in a few minutes to grab a snack or grab something and get on the flight. I think that's a classic example.

>> So how do I use Hudson News? Do I have to download something? I go by that all the time in JFK.

Rajiv Chopra

>> No. We are not in every Hudson News store right now.

>> Oh, okay.

Rajiv Chopra

>> But in the few that we are there, we increasing the number of stores that have Just Walk Out. It's as simple as you tap your card, walk in, just walk out. The example I was going to give is what we worked with Hudson News was on the flow in a couple of these locations, it was far easier for the traffic flow to not put a gate at the entrance, but let that be customers and the shoppers are just coming in, they shop, and then they pay at the exit. So it's a flip of the, you just walk in, shop, and you just tap and leave. So I think that change drove almost 60% increase in revenue.

>> It lures people in.

Rajiv Chopra

>> It lures people in.

>> Inviting.

Rajiv Chopra

>> Exactly. So we do learn from our customers, get feedback from them. And our portfolio of capabilities, whether it's the different types of gates, whether it's how we can install the cameras and the fixtures, that allows for different solutions.

>> I was talking with Julia White who's the VP of Marketing for AWS. She was just on the CUBE. And we were riffing on the concept of a digital twin and how it's not just a manufacturing, it's an efficiency thing. Obviously, the blending of physical and digital as a first party citizen data-wise, both people and their relationships in digital and physical, is really kind of where you're crossing over. You're going to have the digital payments, physical goods, but there's other opportunities beyond that maybe with digital. What's your vision of... Is there a Just Walk In option? Once you're in the system, what are you guys, what's the connection? How do you see the scale coming, if you had to think about connecting the dots, what's next? What's the vision?

Rajiv Chopra

>> I think the vision is obviously to make the deployment of a Just Walk Out store be super simple, super easy to go into any retail store, any outlet... And there are many more that are coming up, like I mentioned, even a break room in a factory, right? Replacing a vending machine there with simple things, that you get not just goods from a vending machine, but you can get fresh things.

>> It's good for the employees too, in and out, get what you need.

Rajiv Chopra

>> It's awesome. Exactly. The increased selection, more convenience.

>> What's the hurdle in your mind, based on your experience, for someone to know, "My environment meets the threshold to go through the process"? Is there like a tool on the front end I would use as a customer? Because it's worth the investment, but you've got to know the environment.

Rajiv Chopra

>> Right.

>> Is there a use case where you say, is there a checklist to say, "Okay, this is really perfect for you"?

Rajiv Chopra

>> Yeah. I think our sales team is empowered with such tools to work with the customers to help them figure out do they actually... Like I mentioned, this is a solution that's great for if you have a high throughput issue. If you want to extend the opening hours or you want unintended delivery or you have a huge theft problem with your other self-checkout solutions, then in those cases Just Walk Out is a great-

>> Solves the theft problem.

Rajiv Chopra

>> And we are constantly, my team is constantly working on reducing the complexity of the product, the cost of the product, the time to launch the product.

>> Is this tied to my Amazon Prime account?

Rajiv Chopra

>> No. Just Walk Out is entirely, that's why it's an AWS product. If you choose to have Amazon handle your payments, absolutely have the opportunity to do so. But if you choose to have completely separate payment infrastructure, nothing tied with Amazon accounts, then our business customers or retailers are more than welcome to.

>> What's the feedback? Can you give me some feedback from some happy customers?

Rajiv Chopra

>> Absolutely, absolutely. I think across different verticals, if I look at, let's say a university scenario, UCSD, University of California, San Diego. They came to re:Invent last year and presented. And I had just started in Just Walk Out at that time. And I was like, "Wow, I love it when our customers are speaking so glowingly about us."

>> And what was the feedback they gave?

Rajiv Chopra

>> They were like, well, it's not just students, but they had a very large staff and faculty that lives on campus, and giving them access to fresh foods. They said their top seller in a Just Walk Out store is fresh salmon; not cooked salmon, but fresh salmon. It's not just a cafeteria, but it's a full convenience store that made the number of times people would come in and shop rather than going off campus, maybe you have a security or safety concern there, convenience concern. So they found like an 83% improvement in their staff utilization, the loss reduction. The shrink came down dramatically. Revenue went up, repeat purchases came up. So I think they expanded from one to four stores of that large size there. If you go on the other extreme, I think I gave the example of Hudson News. I would love to give an example from sports, which is our fans love it. If I look at right here in Seattle, Lumen Field, the Seahawks. They have now expanded eight or nine different stores with different formats, and they've seen like a 47% increase in revenue. And the customer satisfaction.

>> Making it easy is always a good formula for success.

Rajiv Chopra

>> Totally, totally. It's not just that, though. It's a-

>> Business benefit.

Rajiv Chopra

>> Total win-win. The business benefit, the ratings for the fan satisfaction among stadiums, it went from being in the middle to near the top ever since they launched with Just Walk Out.

>> Rajiv, great to have you on, love this segment. What's the vision? What's your future, the rest of the year look like? Share the vision for Just Walk Out, and then what's your goals for the rest of the year going into re:Invent?

Rajiv Chopra

>> I think I am super excited about the entire idea of visual reasoning. Just Walk Out is one example of it. As we think about how we can take the concepts of visual reasoning, just the same way as we are looking at all of the other generative AI capabilities, working with our partners, working with our customers, and adding in more value, providing a differentiated experience that's more intelligent, that's more focused on delivering a magical experience both for the shopper as well as for our business customers, that's my top goal for the rest of the year.

>> All right. Well, we'll be keeping track. Love this Just Walk Out technology. Again, computer vision, it's really tied into the generative AI wave, again, one of the best use cases. We hear multimodal, one of the modes is video and computer vision, pictures. Of course, theCUBE's bringing you all the action here as part of our Halftime Coverage. I'm John Furrier, host of theCUBE. Thanks for watching.