Join RETHINK Retail for an exclusive discussion on autonomous retail!
We take a look at this innovative technology from three key angles: an analyst, a retailer, and a technologist.
Joining this discussion are Sucharita Kodali, vice president and principal analyst for Forrester; Steve Gu, co-founder and CEO of AiFi, a company leveraging computer vision and AI to offer checkout-free shopping; and Mike Fogarty, founder and CEO of Choice Market, a grocery store that recently deployed autonomous technology in its flagship location.
TRANSCRIPT
Julia Raymond:
Thank you for joining us today. I’m Julia with RETHINK Retail and we are going to talk about a very innovative technology, autonomous retail. And we have a view from three different angles today with us. Steve Gu, co-founder and CEO of AiFi, a company leveraging computer vision and AI to offer checkout free or frictionless shopping. Mike Fogarty, the founder and CEO of Choice Market, and it’s a grocery store. They recently employed this feature in their new location in Denver. It’s very cool. If you haven’t seen it, make sure to make a stop and check it out.
Julia Raymond:
And of course, Sucharita Kodali, Vice President and Principal Analyst at Forrester. She’s going to bring some industry insights to this conversation. It’s going to be very interesting with all three of us on the call today. Well, four. An analyst, a retailer, and a technologist. So a lot of different views. And it’s the perfect time for this technology. Although there is a current labor shortage, we’re hoping that that will go away soon. And with that, AI and people will be working alongside each other. So Sucharita, how will this look when they’re sharing tasks with artificial intelligence?
Sucharita Kodali:
Well, there are, I think a lot of ways for technology to help augment the human experience. I think that what we want to be careful about is the notion that, “Oh, this is about replacing labor or that this is about kind of eliminating labor.” And that’s actually not the case at all. In fact, over the last couple of decades, whenever we’ve seen technology come into stores, it’s often to take away some of the tedious tasks and empower those store associates to be able to focus on higher value activity, whether it is managing replenishment, the look and feel of the store, or handling issues that may be customer service related, that can’t necessarily be automated. So in that way, it really is about that one-two punch and being able to get the best out of the store associates that are there, that you’re able to find and retain.
Julia Raymond:
Excellent. So Sucharita, you said it’s not really eliminating labor. Mike, what’s your take on this from the retail side?
Mike Fogarty:
Yeah, I mean, I think retailers now more than ever are facing a lot of challenges across labor, product inflation, and just staffing in general. So I agree that this technology certainly will help supplement and empower our staff to really focus on what we call value added activities. And certainly by employing things around frictionless checkout or even other types of AI, whether that be product recommendation, et cetera, that ultimately allows our staff to focus on creating high quality food, providing the highest quality customer service, and really just providing the customers the best possible experience. And especially in our scenario, we employed a hybrid shopping experience, which allows the customer to choose how they want to check out. And so certainly our stores are not “cashierless”. They very much are staffed by a team of folks that are providing the highest level of customer experience. So I agree wholeheartedly that it will augment and help support our staff, but not replace them altogether for sure.
Julia Raymond:
That’s so important because we’ve been hearing about customer service and customer experience for years now. And it’s applicable in all retail environments, including grocery. Besides the obvious benefits of a frictionless shopping experience, where it’s just so easy for the consumer to grab and go, Steve… Actually let me resay that. Because Grabandgo is your competitor. Besides the obvious benefits of the checkout free, frictionless experience, what are some other maybe not so obvious benefits that AI can provide us? Steve, I’ll pass this to you as the technologist on the call.
Steve Gu:
Sure. I think convenience and the greater user experience is of course the key for success. But I want to also highlight three additional benefits from AI. The first is about more store hours. So with the autonomous store platform, we can now make both the store to run 24/7, non-stop. Historically even today, it’s actually very hard to run a store nonstop because you have to worry about the shifts and the coordination, a lot of efforts going into that space. But whereas the AI and the modern technologies, the stall can run almost by itself. So we actually have early success with some of our clients that the off hours and the weekends contributed significantly to their store revenue.
Steve Gu:
That’s one thing. And the second is about more traffic. So we actually just launched an autonomous store with Verizon partnership with Penske at Indy 500. And that particular day, there are 130,000 people going through the stadium. Imagine all those people going through that shop. Within a single hour, we witnessed hundreds of people going through that store. And it’s simply impossible for a traditional shop, whether using self checkout or cashier checkout to process that volume of transactions. But the autonomous stores can. So that’s another great benefit of handling that amount of crazy traffic within a short time window.
Steve Gu:
And the third is about the data platform. We’ve been talking. We’re seeing tremendous amount of growth on the e-commerce side, but the physical shopping remains as the dominant form of sales in the foreseeable future. What we are doing essentially is to use AI to bring that e-commerce level of convenience and efficiency back to the physical world and to make a tense better. Imagine this autonomous store where you have hundreds of cameras producing trillions of pixels every second. And that amount of information is stunning. So with that, you can understand precisely who is grabbing what, where, when, and how. So you can use that to give retailers very meaningful feedback about inventory, user preference, about stockouts and potentially personalization tailored to customers as well. So in my view, just to summarize about more hours, more traffic, and more data.
Julia Raymond:
Excellent. And with the more traffic, as you said, the Indianapolis 500, the amazing nano store that you guys have there, just handling tons of people coming in and out in a quick burst. Sucharita, from the analyst side, what other retail applications can you see this working well for? How are you viewing this technology?
Sucharita Kodali:
The idea of a nano store, I think is brilliant for every retailer across every sector. It’s a fantastic way to create a pop-up environment and to have it be relatively cost-effective. It can be great for an event like a sporting event, like the one that Steve described. It could be an environment of selling either consumables or school supplies when a university opens, or any place where people are congregating for a short period of time, it makes a ton of sense. It is also, I mean, we’ve seen great applications of these kinds of stores, even in remote locations, for instance, or food deserts, or places where it may not make a ton of sense to staff a complete store, but as Steve had mentioned, there’s that 24/7 capability. And if you can take advantage of all of the information you know about what sells or what’s needed, it provides a great solution for that end goal. It’s much bigger than a vending machine. And it has that ability to be relocated as necessary. And there’s a component, which is really attractive as well.
Julia Raymond:
Absolutely. And you said wherever there’s a large congregation of people, this makes a lot of sense. I’ve heard airports being thrown out there, hotels, things of that nature.
Sucharita Kodali:
Absolutely.
Julia Raymond:
And obviously Choice Market with Mike. I want to ask you because your store is a hybrid. So customers have the choice. They can use the traditional checkout method or use the technology. I know when I visited, I saw both going on. So what would you say is the general appetite for this? How are consumers responding?
Mike Fogarty:
Yeah, I think in the three short months we’ve been open, we’re seeing a lot of adoption. Initially we were getting about 20, 25 orders a day going through Choice: Now. Now we’re getting up into the high forties and fifties. Adoption’s there, the basket size is growing and repeat customers are returning and using Choice: Now sometimes three, four times a week, especially given this particular location, which is at the bottom floor of a luxury apartment complex. So we have folks that are sometimes in there multiple times a day and treat it as their pantry.
Mike Fogarty:
So no doubt, we’re seeing the adoption that we anticipated. But we’re also looking at other formats that could potentially be fully cashierless like the nano store, even smaller footprint, micro marts, et cetera, whether it be as Sucharita said, in an airport, or even the food deserts is something we’re looking at very carefully. Because if we can keep our labor down, our occupancy down, those are two of our largest cost drivers. Ultimately we can then price these groceries for the community more aggressively. And so for us, we think it’s a great solution for food deserts and something that we’re actively pursuing. So I think overall we’re pretty excited about the early data that we’re seeing around adoption and basket size, et cetera. And excited to look at new formats.
Mike Fogarty:
… and basket size, et cetera, and excited to look at new formats, as well.
Steve Gu:
Well, I just want to add to what Mike said, that I think it’s really encouraging to hear that we saw the basket size increase and this amazing customer adoption from choice market. So, that’s really exciting, Mike.
Steve Gu:
I also want to I want to highlight that, although we talk about a hybrid mode, I want to mention that this is actually quite challenging from technical perspective. So, compared to Amazon Go, which is a binary monocular shopping experience where you have to use an app to use a shop, being able to handle both existing customers going through the cashier channels and those app-based users actually poses more challenges for the technology side. And we are very proud of, from our side, that I know our solution is able to handle that complex situation.
Steve Gu:
The fact that the hybrid mode can work is really thanks to our technology advancement, where we are able to configure our solution to be able to use seating camera’s only without depending on additional sensors on smart shelves. And in doing so, we are able to enable this easily retrofit and easily maintenance, as well for Choice Market.
Steve Gu:
And given such, we think that given the early success of the Choice Market with this hybrid mode, we think that you are going to see more and more stores like Choice popup with this hybrid shopping journey.
Julia Raymond:
So, Steve you’re pointing to the flexibility of your solution, but also how it is very complex. And I think that a lot of people might not know, retailers specifically, about all of the solutions out there, the hybrid models, camera only, fully autonomous.
Julia Raymond:
Sucharita, are our retailers aware from what you’re seeing about how autonomous shopping works and what’s out there?
Sucharita Kodali:
There are lots of different ways to approach it. And I think that it’s going to vary depending on the retailer’s budget, the store format that you’re talking about, the type of merchandise. We’ve seen deployments where it can be in a certain part of the store. So, it may be a larger store, but then the autonomous piece is only in a certain area.
Sucharita Kodali:
We’ve seen different approaches to payments where. There’s the ghost store example where it’s triggered by a mobile device. We’ve seen some stores where it’s a credit card that is the identifier. There are some airport examples, for instance, where that is the manner of the transaction.
Sucharita Kodali:
Som there are different approaches that absolutely are being experimented with and tried. And I think that one needs to work through what makes the most sense for both the type of merchandise that you have and what is the comfort level of the shopper with these different approaches.
Julia Raymond:
And Mike, I’ll pass that to you. Sucharita just recapped a lot of good points about that and the different ways consumers can interact with the technology. How was your decision making process going into this before you adopted the hybrid model?
Mike Fogarty:
Yeah, we’ve been looking at this technology for the better part of two to three years, went through a pretty exhaustive sourcing process, and certainly landed with AiFi for a number of different reasons. But we thought that they were the most commercially viable, had some of the leading technology in the space and offered different solutions, not just the fully cashierless model.
Mike Fogarty:
And for us, that was important, especially in a hybrid version because for us, we have a scratch kitchen at all of our stores. So, there’s always going to be employees at the store. And we want it to be a little bit more accessible, so folks can adopt the technology over time as it fits their lifestyle.
Mike Fogarty:
It was really important for us, especially for this store, that we had that option and the customers have that option. And we want somebody to be able to come in and pay for an apple with a dollar bill in this particular format.
Mike Fogarty:
But again, overall, we’re looking at a couple of different formats and sizes that meets the needs of certain applications. AIFI really provides the whole suite of options, whether that be app, credit card in terms of the authorization, but also just these different formats.
Mike Fogarty:
So, yeah, certainly we landed on the hybrid version for this particular store and are really excited about the adoption that we’re seeing.
Julia Raymond:
That’s great, Mike. And you mentioned the commercial viability of AIFI was at the top of your list and the ease of deployment, affordability, things like that. Those are all amazing points to help implement a hybrid solution.
Julia Raymond:
And Steve, from the technology side, I wanted to ask you because your technology spans many countries and there’s a lot of variables that come with that. The AI is always learning. Does it track customers differently region to region?
Steve Gu:
Yeah, that’s a good question.
Steve Gu:
I will say, from day one, we designed our platform, our technology to be country or region agnostic. So, getting the store running is really complex. And well, first of all, we make our solution very modular so that it can adapt to a different store shapes, fixtures, different layouts. And we are very successful doing that.
Steve Gu:
That’s why you see that we’ve launched many stores in many different countries. Our technology can recognize products and people no matter where they are, no matter which country they come from.
Steve Gu:
Second is more technical.
Steve Gu:
In order for the AI to work, we actually developed a very unique approach using simulation and the insight from the data sets to empower the AI to learn by itself without resorting to very labor-intensive, manually labeled datasets.
Steve Gu:
So, through our simulation platform, we are able to simulate to all sorts of different user behavior: your hair style, your facial expression, your pose, your clothes. We are able to enumerate thousands and millions of store configurations just entirely virtually without the need that to do those labor intensive jobs of adjusting cameras in the physical space.
Steve Gu:
With that, we make our solution very scalable to different store environments.
Steve Gu:
And if there is a one particular story environment that our solution can not handle, we can easily put that into our simulation environment and train the AI to learn about that and therefore to accommodate for that particular store environment.
Steve Gu:
The third is about privacy.
Steve Gu:
So, we have designed our solution from day one to be very sensitive about customer data. Our solution is GDPR and CCPA compliant, so that our customers can launch the store to the public with this piece in mind. For example, we have launched several stores with our clients in Europe. They are GDPR compliant. They don’t use a facial recognition or biometrics. And this way, we only collect anonymized data so that our solution can handle multiple situations without violating any user privacy concerns
Julia Raymond:
And Steve, that’s huge. That’s absolutely huge for all of the companies based in Europe, all the companies who operate in Europe. The fact that you guys have synthetic data that can create the simulations creates a lot of benefits, including the ability to scale faster and make it more profitable.
Julia Raymond:
Sucharita, from an industry analyst point of view, what’s your take on the future of nano-stores?
Sucharita Kodali:
The idea of these smaller autonomous, semi-autonomous, could be fully autonomous stores really, as I’d mentioned earlier, just makes complete sense for everyone. It’s a great way to have a broader store strategy, to have kind of a store strategy that fits the needs of shoppers, without having to necessarily have to invest in long-term leases or be stuck with a larger store footprint than what you would be stuck with if it was a traditional real estate model. So the idea is actually a brilliant one. And this idea of autonomousness in this format makes sense because, as we know, there can be a lot of variable cost challenges and you would limit the number of hours that potentially a store like this could be open, which really provides a lot of great ability to have flexibility, to test and learn and to ultimately serve consumer needs wherever they may be.
Sucharita Kodali:
So I think that there are some really, really interesting reasons to be experimenting with this. This, I think, is a fantastic add-on to whatever a store strategy may be. This is never going to be something that’s going to completely replace a large format store or a major grocery store. I mean, there’s only so much that a nano store can hold. But the idea of being able to take it to shoppers wherever they are and provide them some really useful products that are fast-moving consumer goods, that can be really, really useful. And as kind of we talked about earlier, the use cases often seem to fall toward grocery and convenience because that’s where the fastest moving goods are, but there are a lot of really interesting use cases as we’d alluded to earlier with sporting events or specialty scenarios where it could be about selling apparel or licensed merchandise or some limited edition types of products that you may be looking to showcase to an audience in a very limited fashion.
Sucharita Kodali:
So I think that there are a lot of really fascinating use cases and we’re just scratching the surface of what all of it could be. But I think that brands have a lot of room to experiment here. Retailers that are often known have a lot to potentially experiment with, and even new retailers that are looking to touch more consumers or have consumers engage with their brand, but don’t necessarily want to sign those 5, 10, 15 year-long leases, this is a great way to get there.
Julia Raymond:
Absolutely, Sucharita. And you made me think of something because you said hours, extending hours, is a big thing for retailers. And we saw in the past, some huge retailers that were typically open the holiday closing their store hours. But if they adopted the format you had mentioned earlier with certain sections of the store having this technology, you could get a present last minute and not have to rely on our supply chain and the delivery guys working overtime. So that’s amazing as well.
Julia Raymond:
Steve, from the technologist side, taking a step back really quick, what about you? Where do you see the future for this?
Steve Gu:
Yeah. Well, from a technology side, I think the future is already here. It’s just not evenly distributed. In fact, I think in Choice, we are already seeing the future of retail. So in my view, there are three different trends we’re seeing. The first is about convenience. I think Choice is a great example here where people can just walk in, grab stuff and then leave. Everything is already automatic. And the nice thing about Choice is that it’s inside apartment building. So literally people can just walk downstairs and grab stuff. And that only takes five minutes to finish their entire shopping trip. This is faster than the fastest delivery options today. So I think that this actually fits very well with the trend that instead of asking people to come to a central place, to drive miles away to a central place to shop, I think in the future, you’re going to see more and more stores located as close as possible where people live and thereby bringing supreme convenience to the local neighborhoods.
Steve Gu:
The second part is about the user experience. Again, I want to use Choice as an example, because I really love Choice. I think that offers this very nice, relaxing shopping environment. It also, like its name Choice, it also gives choices back to the customers. So the customers can choose whichever way they want to shop. We think definitely you’re going to see those hybrid model shopping journey happening in more and more shops like Choice does today.
Steve Gu:
So actually my background shows there’s this little nano-stall sitting there. I think Michael has already mentioned about nano-stores and potentially its use in food desert. Basically it enables this AI and these autonomous platforms enable new store formats. So these were concepts that almost impossible in the past. So nano-store is a great example. So with nano-stores, you can place those micro-shops almost anywhere from universities to stadiums to concerts and to place those stores where people want them most. We think that potentially in the future, you’re going to see those automated nano-stores popping up in almost everywhere, like a mushroom. So in the next 10 years, we think that this autonomous store concept, this autonomous store platform, will transform from a novelty concept to something almost essential that every retailer needs to equip with.
Julia Raymond:
Absolutely. So you touched on supreme convenience, great user experience and automated stores becoming ubiquitous. Steve, I’ll pass this to you as well, you work with a lot of retailers at AiFi, what are you hearing? What do they want to see more of?
Steve Gu:
According to our interaction with many retailers, they definitely viewed as AI, that’s autonomous retail as one of their key strategies moving forward. One thing in particular is that, despite this rapid growth from e-commerce side, so most of the retailers believe that the physical retail will stay strong and thriving, especially for those high frequency, high margin essential and convenience goods. And traditionally you can do a lot of interesting things online with e-commerce website, you can measure everything of who you are, what do you grab.
Steve Gu:
But now with technology and AI, we can do essentially the same thing in the offline world as well. So with our AI platform, we can literally turn those physical shopping grounds into the equivalent of e-commerce websites. I think that represents a fundamental shift towards this digital era. And second is that, for the retailers, we think that they need a winning strategy. How to leverage AI in autonomous stores in a way that can really help those stores stay strong and thrive. There’s a lot that we can do, but fundamentally it’s about how to get more goods in front of more customers, in more locations and with more choices. And thereby bringing more fun and engagement, and driving customer satisfaction.
Julia Raymond:
Steve, that is such a huge point. The fundamental shift you mentioned, being able to basically have the same amount of data, if not more than you would get from an e-commerce site. And be able to be agile and change things, and give the consumer what they want, which is so difficult. It’s increasingly difficult. And we’ve heard from an analyst and a technologist. So I will pass this to you, Mike, as the retailer on this call, what do you see? What do you want more of?
Mike Fogarty:
Yeah, I think Sucharita and Steve did a good job outlining some of the areas that retailers are looking at in terms of deploying AI, and specifically autonomous stores and frictionless checkout etc. I think ultimately we want what the customer wants, because as Sucharita said, we want to drive revenue, we want to reduce our costs and be a profitable and sustainable business that can grow and scale both here in Denver and throughout the country. So these technologies certainly help in that regard. And I think being able to better leverage that data as Steve said, there is a significant amount of data. Things like dwell time and gender, and things of that nature, up until now I haven’t had that insight into who’s actually coming into the store.
Mike Fogarty:
Yes, you can measure that on an e-commerce platform. But now as Steve mentioned, we can measure some of those things in the store. Things like heat maps of what parts of the store is performing the best versus others. And then we can drive merchandising decisions and things of that nature. So I think overall, really leveraging the full power of platforms such as AiFi, and working close with them really to develop these features and functionality because it’s mutually beneficial for both of us to really get down into the analytics and the data that these systems provide. So I think we’ll be working closely together on building out that suite of features and functionality that obviously it can benefit choice, but also some of their other customers as well, so.
Julia Raymond:
Certainly. And Steve, I’m going to mention something you said earlier, the future is already here. And Sucharita from an analyst perspective, what is the biggest indicator to you that it is here? That autonomous tech is taking off?
Julia Raymond:
Certainly. And Steve, I’m going to mention something you said earlier, the future is already here. And Sucharita from an analyst perspective, what is the biggest indicator to you that it is here? That autonomous tech is taking off?
Sucharita Kodali:
I think that there were a lot of questions back when, say for instance the very first stores came out, like Amazon Go, around the cost and around the viability. I think what Mike has demonstrated is that, there are absolutely reasonable and viable approaches that companies of all sizes can consider. There are different use cases and examples, and there are, of course, the food deserts, he mentioned the apartment complex. Which are very different use cases. But with the right assortment and with the right technology, which has developed and evolved to a point where it is much more affordable and it makes a lot more sense for experimentation, and it offers an opportunity for innovation. And also I think that where we go next from here is probably also a bit of experimentation with the retail industry.
Sucharita Kodali:
Overall, one of the biggest things that we’ve been hearing about is the idea of retailer media networks and retailers bringing in different types of advertising, creative advertising, leveraging advertising as a means of driving profitability at a store level while also serving brands and providing value to shoppers. And an autonomous store is a great format to pilot and test some of those types of initiatives. Brands are constantly looking for new ways to get in front of consumers, particularly for new and innovative products.
Sucharita Kodali:
And what better way to do this than to do it in a smaller format store in the beginning versus trying to do a national rollout where it’s a much larger investment. So I’m excited about the possibilities that exist and the fact that the price points now seem much more reasonable. The fact that there are deployments with different types of retailers, in different types of settings. That the only use case is not a big company out of Seattle, it’s really promising. And I think that there’s a lot of exciting opportunities for the entire industry to consider.
Julia Raymond:
And Mike hats off to you, Sucharita mentioned you are proof that this technology is taking off, that it is within reach for all retailers. And so we appreciate people like you helping move the industry forward. And Steve, Sucharita, Mike, thank you all for joining today. This was a great chat.
Steve Gu:
Yeah. Likewise.
Mike Fogarty:
Thanks for having us, yeah appreciate it.
Sucharita Kodali:
Thanks Julia.