EP 218: The Next Frontier: Agentic AI Smart Shopping

Guests: Bridget Johns, co-founder and CEO at To&From, John Andrews co-founder and CEO of Cimulate.

The future of retail is being fundamentally transformed by artificial intelligence, moving far beyond basic search functionality to create intuitive, personalized shopping experiences. Learn how AI is evolving from simple keyword matching to sophisticated understanding of consumer context and intent. The key to success lies in striking the perfect balance between AI automation and human expertise. Join Shelley, Bridget and John as they discuss how the emergence of Agentic AI -- autonomous shopping assistants that can understand complex requirements and preferences -- represents the next frontier in retail technology. The human element will remain crucial, particularly when deep product knowledge and curation are required. The most successful implementations will combine AI's processing power with retail domain expertise to enhance rather than replace the human touch in shopping experiences.

For more strategic insights and compelling content, visit TheRobinReport.com, where you can read, watch, and listen to content from Robin Lewis and other retail industry experts, and be sure to follow us on LinkedIn and Twitter.

Transcript

Transcript by Descript: 

 And this conversation can kind of help engage in a way that doesn't feel like a chatbot. It most certainly can't be a chatbot. It needs to be more engaging, but provide more of a shopping assistant where that shopping assistant has the brain of understanding everything about the products and ideally understanding everything about the customer. 

Retail Unwrapped is a weekly podcast hosted by Shelley Kohan from The Robin Report. Each episode dives into the latest trends and developments in the retail industry. Join them as they discuss interesting topics and interview industry leaders, keeping you in the loop with everything retail.  

Hi everybody. And thanks for joining our weekly podcast. I'm Shelley Kohan, and I'm very excited to welcome Bridget Johns, co founder, CEO of To & From, and John Andrews, co founder and CEO of Cimulate. So, I am thrilled to have you both on Retail Unwrapped. We're going to be discussing the role of AI from a consumer shopping perspective.

And I think before we get started, I'd love for our listeners to know a little bit more about To & From and Cimulate. So Bridget, if you don't mind, I'm going to start with you. And before I turn it over to you to tell us about the role of AI in gifting, specifically. I want to mention your background is quite awesome and impressive.

You come from retail links of London, Lancôme, Tiffany.  and then you kind of morphed into this retail tech space, which then led to you starting your own business, which is really linking retail and tech. So tell us about To & From.  Yeah, thank you, Shelley, for having me. John, great to see you.  I'm excited to talk about To & From.

It's the gifting season, so I think it's like an apropos time to have the conversation. Retail tech  To & From is really a retail tech company focused on building better gifting experiences that live throughout the digital commerce experience. And when you talk about my background, it's like  plays right into how we think about where the vertical AI will win.

And we think of gifting as a vertical AI application that really needs to live everywhere consumers are. And we think that the winners. Will be people who come to the space with really deep domain expertise. So I am one of the retailers on our team. We have many others.  and we really think about the problems around gifting from both the retail perspective, but as well as bringing the technology and to support that. 

That's great. And, we're actually going to talk more about that later on about the retail expertise. So we're going to take a little deeper dive in a minute. And then, John, first of all, you're so modest, but you actually co founded another company called Select, that was bought out by this little unknown athletic company called, what was the name?

Oh, Nike.  and your background's interesting. You were actually schooled in computer science and you have an MBA from Harvard.  so you kind of come from the tech product side of the business. So tell us about Cimulate. And because I'm such a super curious person, why do all your companies start with C? 

Yeah, yeah, we suspicion suspicious at this point or,  we are,  we, we, we like the soft C,  and since it worked for us at select, we were looking for options at Cimulate that that worked as well. And that whole idea is, it's all about the customer,  for Cimulate. It's a customer simulation is kind of where we started at select.

It was a, it was a customer election.  and so we,  founded this company founded Cimulate. I founded it with my co founder Vivek Barias, who's an MIT professor. He actually co founded select.  With another M. I. T. Professor and I joined on super early on as CEO. So I wasn't technically a co founder there, but I was in the very, very early days.

Yeah, we had had a great run there. you know, some really interesting A. I. Focus technology. And then sold it to Nike, and it was a really was a great marriage at Nike, and I spent about three and a half, three and a half years working, working there. What we're doing at Cimulate is actually driving a, driving a generative AI experience focused on commerce and using large language models.

To do that. And the whole idea is, is, is if you can understand at a much deeper level, the customer context and what the customer is looking for, and if you can understand the products at a much deeper level and kind of bring that context in terms of understanding then what content to deliver to customers through that customer experience, and you can do that in a more kind of personalized manner. 

Obviously you would want to personalization has been something that we've talked about in retail for years.  how do you optimize the experience? How do you deliver more relevant experiences? Prior to select, I ran the commerce product at Oracle and came into Oracle via the acquisition of a company called and DECA that did search and it Cimulate.

We're focused on search. I did not think I would be back in search.  But we are, you know, there's a really big opportunity right now with the new technology stack, and we're pretty excited to go after it.  Well, that's great. So that's a nice segue into our first topic, which both of you are going to tell us a little bit about how generative AI is going to absolutely change the consumer shopping experience.

So, John, let's start with you on this.  Yeah, so this, this question comes up pretty regularly just because of all the noise. around generative AI. and so, you know, AI has been around, AI has been around for a long time. Select was very much more kind of machine learning, some really interesting algorithmic approaches, tensor completion, right?

Ways of understanding sparse data to make better, you know, make better decisions in, in, in retail, but generative AI, and particularly You know, transformer models, large language models really is a new generation of of technology.  and the question comes up is like, hey, is the hype for real? Right?

Because it is. It is a bit definite. And and. You know, like any good generative AI startup founder, I'm running around with my hair on fire saying it is, but I actually really do believe it is. And in some cases, and particularly within commerce, it might be under hyped a bit, right? Just because I do think it is going to have a pretty significant, a pretty significant impact.

This idea of understanding the context and then how you use language In terms of communicating and how you can use these transformer models and a one way to think about it is,  if you played around with check GPT and you kind of see, you know what it, you know what it can do. Claude Gemini, whatever, right?

These large language models are effectively a sequence model. Okay. And the sequence model is basically next best word. It is mathematically understanding what that next best word is likely to be, but it doesn't in a way. It'll just kind of blow your mind, right? What we're doing is using that same kind of transformer model technology to do next best action.

Okay. And so you can understand the sequence and you can think of a prompt that you would put into chat TPT. Think of the prompt that you would put into a commerce focused large language model with all of the prompt being understanding the customer's history and everything that they've done and saying, what is the right set of technology or the right set of products to put in front of this customer?

And I think that this, you know, this generative AI tech is actually going to play a pretty significant role. Cool.  in terms of over the next couple of years, many technology, many retailers needing to replatform their stack to leverage more of this technology within it. And and obviously with just starting a company in this space, we obviously hope that's the case.

So take it with a grain of salt coming from the founder. But I do think there's a pretty big opportunity here.  That's great. I'm Bridget. I'm sure you want to weigh in here, too.  Yeah, I mean, we think about the problem in a similar way is what John has described. But the thing that we. Come back to time and time again, having  been sort of exploring the space and trying to figure out where best to support consumers and their gifting journeys is really.

Who controls that end experience and will the sort of riches go to the incumbents or will it be a totally new experience that consumers have and we're sort of taking the road that gifting lives everywhere. So it can live in a B2B experience. It can live in a consumer experience. It can live on a marketplace website.

It can live on just a single brand consumer direct to consumer website. And we think that you have to be able to drive the right experiences everywhere. And part of that is, as John says, this, you know, being able to understand natural language and being able to support that with the right gift recommendations.

But there's also a whole other set of. Tools that you need to really do gifting well that,  also can be like very well supported with AI. So things like reminders or even just a very simple calendar reminder that pops up two weeks before your wife's birthday to say, Hey, John, your wife's birthday is in two weeks.

Have you done something about it? Here's a list of gifts that she might enjoy.  so. So I think that,  AI is going to like transform, not just, and I think we're going to talk about this in a little bit, but not just like the recommendation piece of it, but in fact the entire experience, especially when you're, when we're, we're talking about gifting. 

I love that. And you're right. I think both of you are right. I, there's so much like hype about generative AI and what it's going to do that, I think retailers are a bit confused to be honest about how they can actually. You know, put something together that makes sense and can actually impact the shopping experience.

So I've heard in the industry and I, you know, as a retail pundit, I write a lot about this quote unquote, conversational commerce and how AI is going to change the shopping experience through conversational prompts.  love to hear your thoughts about that.  Maybe John, you can start.  Sure. I'll go, I'll go quick.

And I know, Oprah just got some perspective on this as well, and her work attuned from.  I do think the, I do think one of the benefits and why I think, like, generative AI and large language models are under hyped a bit is because of what is possible from a conversational perspective.  you know, if you think about how and and I'll just let's let's start with search, right?

Because I do think when you think about the conversation with a customer online, the back and forth, you can think about it as kind of a query and response every time somebody comes to a digital website, they're going to click on something. That click is on either a category page, the search box, an offer,  and, you know, a sale item, a product detail page.

It's giving context of what the customer is looking for, and then the retailer needs to return back kind of a set, you know, a set of content. And search is that foundational kind of query and response. You're querying the engine, saying, here's the context. Bring back the right products and order them appropriately. 

And historically keyword search has been what's ruled the day, right? It's been hard to beat for 25 years. Hopefully the customer in Google has trained us to do this in a way where try to think what the most important words are and put that in the search box. And hopefully you get lucky right to the point where Google even early on had the like, I don't even know if they still have it to get lucky, but  right where it's just, it'll just, it'll just give you kind of the top result. 

Then there was AI search, which honestly was more. It's vector based search, right? It's not large language, not LLM based search.  and what is possible now with an LLM kind of operating system powering this customer experience. Is that it can be more conversational, right? I think there's a way where you, you optimize the recall of products, but you can ask questions that are rather than like sweater or merino wool sweater.

I can say, I like thin sweaters, crew neck sweaters that are, you know, merino wool or something like it. What do you have? Right? And I can put, you can put that in. And the, the, the retailer should be able to understand that query, bring back the right results, and also come back with other additional prompts that might be, you know, that might, might be interesting.

Is there a particular brand that you're, you know, that you're interested in? Do you like a design or just a solid color? And this conversation can kind of help engage in a way that doesn't feel like a chat bott it most certainly can't be a chat bot. It needs to be one or more engaging, but provide more of a shopping assistant where that shopping assistant has the brain of.

Understanding everything about the products and ideally understanding everything about the customer. So some of my earlier points.  Love that. Yeah. And I, yeah. And I think like, I agree to a point, but when we think about conversational commerce, we have to, I think, really focus on where that conversation happens.

And a lot of shopping happens when you're on the train or. When you're like waiting for your kids to come out of school or whatever it is and you don't want to be talking But you also don't want to be endlessly typing So I think there's this like There's this friction that exists because if you could just talk, it would be way easier.

But when you get into this conversational commerce and that conversation has to be like endless keyboarding, that is not a good experience. So I think like being able to anticipate finding the right user experience to like fast track that understanding. More about who you are as a person, so you can sort of anticipate what you would say and,  and help to fast track that.

I think it's going to be super important because I do think people are going to have fatigue from having to type into these conversations when really what they want to do is just like,  Not, not say a spoken word.  Yeah. If that makes sense. I don't know how you're thinking about that, John, but that's something that like, I think about a lot.

I'm like, nobody wants to be typing nonstop. Yeah. That is not fun. I, I, I complete, I completely agree. And even just if you do speech to text, right. You don't wanna be on the chair, on the train, always gonna talk into your phone. Right. So, so completely agree. And I think this is the one, this is the thing.

Where I think it's gonna get super interesting in the next couple of years to see what the customer shopping experience looks like from a digital perspective, right? Because I think historically being able to refine by attributes and size, color, price, that's not gonna that's not gonna go anywhere. I think that's still an important element to quickly refine along if you have a larger catalog, right?

But you should be able to put in perhaps more descriptive queries Where you're not just putting in keywords and hoping those keywords match, but what the, what the conversational agent needs to be able to do what that, what that, that engine needs to be able to do is exactly what you say, Bridget, bring back suggested responses, right?

To say, Hey, are you looking for a particular, you know, brand and then show the brands and, and, and do it in a way where the customer can then click. But then also refined by attribute. It's not it is not a chat bot, right? It is not this like little clippy in the lower right hand corner of the screen where it's a back and forth and said, like, oh, here's one product.

Maybe you'd like and it shows you a picture and send you a link. It needs to be that cohesive experience  that where you don't have. I completely agree. Bridget.  Yeah. And I think also, like, the thing that we think about with gifting, which is, you know, It's even more complicated than search because you really have, you have three inputs that you have to deeply understand.

So you have to understand the gift giver and what they're looking for and what their preferences are and what brands they're like. Interested in giving as a gift, you have to understand the giftee, like what are they into? And then ultimately you have to understand the brands can support what those two people are looking for.

So when you think about building the AI and the conversation between these three different stakeholders, it's incredibly complicated and, really just really exciting to be working on. So,  I'm hoping that we're, we're going to be the ones to figure it out.  I definitely think both of you are going to be the ones to figure it out because as we were discussing the podcast, I learned something new, which quite frankly, whenever I spend time with Bridget you or John, you, I always learn something new. 

So, but I would love for you to kind of explain to our listeners about agentic AI.  So what is it and how is this going to change how people shop?  This is all you've done.  there's a, there's a ton of research and there's a ton of development going on on the agentic on the agentic. AI,  in the agentic AI landscape right now,  if we just think about first, like, what is what is the definition of it right?

And  try to keep this relatively simple. Right? But it's. It's this idea of using artificial intelligence,  to use a I to autonomously  go out, perform specific tasks with specific instructions and then, you know, come back with some, some, some answers and potentially even perform. Additional tasks, right?

You can think of this in some very simple examples as,  like a personal assistant, right? Where you can set it up to, like, go through your emails and automatically schedule meetings, right? And just kind of it will go out and it it's got an element of workflow into it, right? You give it some instructions and then it will do a, b, c, and d what I think is going to get interesting from a it, right? 

From, from a commerce perspective and from a retail perspective is imagine you had an  agentic shopping  assistant. Okay, where you can say, hey, I'm looking for, a new travel bag. Right?  and. My my agentic assistant, I would probably have a better name for it.  go out and look. I'm looking. I want something that, I don't like roller bags.

I like duffel bags, but it has to have a really comfortable shoulder strap that I can wear on a messenger bag. it needs to be at least, I don't know, 40 cubic something or other,  and it needs to be waterproof. And I like neutral colors and from a cool brand.  And then it's going to go out and it can go out and it can bring me back and summarize every word of what that it found.

And then I can say, I like this one. And then I can go out and buy it. And get it shipped to my house, like the whole shopping experience. All I care about is describing what I want, maybe making some decisions. You can think about it. Somebody coming into the dressing room, holding you, you explain it, holding up, but then having everything else just automated, automated for you. 

Yeah, I think it's such an interesting space and it's actually one of the,  earliest verticals that we've been working with are companies that are. Building their own personal AI models to support consumers in their everyday lives. And it's interesting because some of them are really going down the agentic AI route, where they really want to do anything, everything automatically and without a human in the loop. 

And some of them are saying, no, no, no, we need a human in the loop to  Validate and to help you make the decisions, but the work on the back end can be with agentic AI. So,  I don't know, like we, we hear about another, a different company working in the space literally every single day. I'm sure you do too, John and,  It's just really interesting to see like who's who's in the race and who's going to win the race because I think there's like Definitely not a clear winner right now  but lots of people trying for sure.  Yeah, I I If I mention I think also, you know, it's it's interesting  when when a lot of people talk about, like, what is the what is the future workforce look like?

What is a I from a jobs perspective? What impact? I think it's where it's the agentic side of it that starts to get interesting. Explore. Well, what tasks are these folks going to do?  and historically, we've seen technology impact blue collar jobs, manufacturing.  like just kind of very wrote over and over, you know, things that can be automated, perhaps with robotics or whatever.

I think what gets interesting is some of the things that can be done. Now, with with a I and some of this workflow,  that can even, you know, impact within the retail space. Right. And, and I think it's gonna be something that we're gonna have to watch and  monitor because I think there's a lot of lower level, like lower level, but skilled jobs within, you know, retail, you know, retail companies that some of these agents are gonna be able to do.

But that's the training ground. For brilliant merchants, right? And planners and allocators and,  you know, store folks and supply chain folks kind of down the road. So I think it's, you know, on the on the other side of this, something that's got to be interesting to see and kind of keep an eye on kind of how it evolves into the workforce as well, not just in the commerce experience. 

Yeah, I think that's so true. And, I'll, I'll comment a couple of things. One, I'd love to have, like, I could have like 30 shopping specialists out there looking for stuff for me. Right. So can't, it doesn't have to be limited to one. So that's exciting to have, you know, all these shoppers out there curating products.

But also what you said about supporting the workforce, I really believe that's a great skill. So if you imagine your. Example of the shopper in a fitting room or a shopping experience, and if that agentic AI can kind of curate for  a personal shopper the best of the best and the most, you know, unique products for that particular customer, and then they have this experience with the customer personally, one on one.

You've just increased their productivity tenfold because they're not running around all over the town looking for different products. They can, you know, very easily curate those products. So I do believe it's going to help, you know, the other functions, the buying function. For example, I'm sure there's not one buyer that loves, you know, inputting purchase orders or doing all kinds of the behind the scenes stuff, or they're planning a buying trip and they have to do all this research about where to go.

What if agentic AI. mapped out the, you know, their entire buying trip. How great would that be? So I do think there's lots of applications to actually support workers in retail and on the brand side.  Yeah. And I just don't think that AI is anywhere close to at least the line of sight that we have. The curation piece is so important and you can't like.

Just say, give me three travel bags that have a good shoulder strap. Cause you're going to get a lot of crap. Like you have to have somebody who cares deeply about the curation, about who it's for, about like all of the different, inputs that are related to style and design that. You just like even as a consumer, you don't even know to ask those things.

So it's not like you're going to give this very robust prompt that says, I want double stitching on the, on the gusset. Like you're not just not going to do that. It's not how consumers work. So the curation piece, I think personally, I think for any of these applications to work still has to be a human led expert led,  piece of the puzzle. 

Well, Bridget, you mentioned that one of the keys to your business success is understanding, you know, who you're buying for and the person making the purchase and understanding those elements is very important. It's the human piece of it. And you're right. No one's going to ask for these kind of jargon words that we used in the industry.

So being able to, create these products and have a human actually understand the buyer really well to be able to, Yeah. go through and select products that make sense for those consumers. Yeah. Yeah. I think having this, sort of curation led  technology I think will be, because I've seen John, some of the work that you guys are doing, it's so interesting and amazing how you can really change the way That your customer sites think about a prompt or a query when somebody is looking for something, but it's already a curated set of product.

When you try to do that in the wider web, it falls apart completely. Completely agree. It's a it's the task is the task is not well defined at all. Yeah, it becomes really, really difficult. We agree. I think the other big problem is consumers don't know what they want. I was looking for a dress and I said, I don't want black.

I don't want a slit in the front. And I ended up buying a black dress with a slit in the front. So, like, you know, we don't even know what we want. So how do you, how are you going to manage that part when consumers really don't know what they're looking for in these? Types of instances with  curation. Well, when I'm gifting, I go to To & From. 

Her team, tell me what some of the best gifts are that people are going to like and that I'll feel good about giving so that it is, it is an important part of the process.  Definitely. So when I listen to both of you, there's one inherent characteristic that you both have that I believe drives your success and the companies you work for and probably really creates your vision for your own companies.

And that is, and I truly believe this. in practice, and that is you both understand the retail experience, your industry experts about retail. So Bridget, I'm going to start with you on this one. And let's kind of circle back to what you mentioned earlier about how the best retail experiences are actually from people who come from the industry. 

Yeah, I mean, I think that, you know, this goes back to my work at retail next when we were, you know, I came out of retail and I was working for this technology company that was really focused on helping brands understand,  data in their brick and mortar stores and really like trying to get data out of those stores to help them optimize what that shopper experience was.

And the, Yeah. Biggest thing I brought to the company was my deep understanding of retail. I didn't know anything about technology when I started in that company, like literally nothing. And,  I had huge value that I brought to the organization because of that domain expertise. Now, of course, like I was there for many years and I learned a ton about technology and really sort of expanded my lens,  added and added to my tool box.

So.  Now I feel very well equipped to solve the problems around gifting because I have both the technology and the retail expertise, but  I don't think that somebody coming to a retail problem without retail expertise, either, you know, as part of their team or even just The intellectual curiosity around it is going to win.

I think it's like the number one thing you have to have. You have to be so deeply focused on the consumer and the problems that they have and how you're going to solve those problems in a way that resonates with them.  I had somebody tell me, I moderated a panel probably 10 years ago with a retailer, and one of the things he said to me has stuck with me.

He said consumers are so far ahead of retailers in terms of their expectations that you have to be so dialed in to what their expectation is, because like the idea that you have is probably not what the consumer is expecting. So you have to be listening and talking and experiencing all of that. All the time to build something that is going to like really solve problems for for everyday consumers  John any comments on this retail expertise behind the scenes  Yeah, no, I do.

I think I mean obviously I think it's important. I think One thing that's interesting is when Vivek and I were, were, were ideating on what the next big thing was going to be, there were a couple of things that we were looking at. This was probably, you know, two, two and a half years ago.  and, you know, we had this one idea that wasn't like directly kind of retail tech related and, We met with one of our, like, mentors, advisors, a guy, a guy named Steve Papa, who was the founder of the company, was the founder of Indeka, and he's the guy who actually introduced Vivek and me to each other,  and, he was like, yeah, but what's your comparative advantage?

Like, what, why you guys, right, doing this? Compared to somebody else doing it. Whereas what we're doing now from a commerce perspective, from a generative AI perspective, from a customer experience, and particularly with some of the VEX research and like customer choice modeling.  it's, the, the, the con, the concept of founder market fit actually feels like as opposed to product market fit, which I feel like we've gotten, I feel like we have  founder market fit  in a way where,  there's, we, we, we, we know the landscape, we know kind of how people think about, we know where some of the traps are. 

But I think what's more exciting for me right now is  I really do want to come back to one of my first comments. I do think the next couple of years is going to be really, really interesting. So the thing I'm excited about, it's just not more of the same, right? It's going to require some new skill sets.

It's going to require us to think a bit uncomfortably about  you know, what, what is possible and how do we, you know, how do we get there? but it's in a space that I know and I love.  That's great. Well, I want to thank you both for being with me today. We certainly learned a lot, but any closing comments you want to add before we say goodbye? 

I mean, I would just be irresponsible if I didn't say it's the end of the holiday season. There's still a few more shopping days left. Make sure you are, being thoughtful in your choices for the people that you, are gifting this holiday season. If you need some help, just let me know.  Great.  Yeah. And it's a great experience.

I'll echo, what, what Bridget said for To & From and some great ideas.  I'm going to go back to my very first comment where you asked me why the C. And as I, as I, after I was saying it, I think I said, we're we're suspicious. We're superstitious.  We are very suspect and suspicious, but we're superstitious in terms of, in terms of the sea and.

I do. I do think it is. It comes back to, like, just understanding the customer, understanding the consumer and delivering experiences that are, you know, that are, that are fun to deliver that give business value.  and, yeah, I'm excited with what we're seeing in a lot of this conversation around AI and generative AI.

What, what's, what's possible.  It's very exciting. So thank you both. And thank you to our listeners. If you have any comments, suggestions, ideas for future podcasts, please reach out on the  rodhamreport. com under contact us. And next week we have a fun review of the best podcast highlights of 2024. And it was really hard getting down to just a few. 

I think we're heading into our 220th podcast, so thank you for all our listeners for your support. Thank you, Bridget and John.  Thank you for listening to Retail Unwrapped. We'll be back in one week with another podcast. Please subscribe on Apple Podcasts, Spotify, or any podcast service. If you have questions, ideas for a podcast, or anything else, please contact us via therobinreport.com.