How Generative AI is Transforming the Retail Experience

In this podcast episode, Dr. Beverly Wright, Vice President of Data Science & AI at Wavicle Data Solutions, sits down with Andrew Simmons, Retail Practice Lead at Wavicle Data Solutions, to explore how generative AI is reshaping the retail landscape. From personalized grocery lists to in-store customer experiences, they discuss practical applications, potential pitfalls, and the long-term vision for AI’s role in retail. This episode offers an engaging look at how generative AI is revolutionizing the way customers shop and interact with brands. 

Speaker details:

• Dr. Beverly Wright, Vice President of Data Science & AI at Wavicle Data Solutions    
• Andrew Simmons, Retail Practice Lead at Wavicle Data Solutions 

Watch the full podcast here or keep scrolling to read a transcript of the discussion between Beverly and Andrew: 

Beverly: Hello. I’m Dr. Beverly Wright and welcome to TAG Data Talk. Andrew Simmons, who is in charge of the Retail/CPG vertical at Wavicle Data Solutions, is with us today. Thanks for being here, Andrew. 

Andrew: Thank you for having me, Beverly.

Beverly: Excellent. We’re talking about the top challenges and opportunities gen AI brings to the retail experience, a topic that’s going to be near and dear to just about everybody. 

Andrew: On one side or the other.  

Beverly: Yeah. But let’s start off with a little background. Tell us, why are you so cool?

Andrew: I’m glad you asked. I’ll tell you what brings me to the podcast studio today. I hope that makes me cool—about 18 or 19 years in the data space. With experience comes a little perspective about the evolution of trends. So, I joined up in the 2008 and 2009 space. Basically, I’ve been in it my whole career. At that time, I’m trying to remember what some of the buzzwords were, big data was definitely hot.  

Beverly: Oh, gosh, I remember that one. That was embarrassing.  

Andrew: Yeah. I remember my very first gig, barely knowing what I was doing. Actually, and I’ll omit the names to spare the innocent, I was staffed as a developer on a particular ETL tool. I had exactly four days of training.  

Beverly: Oh, boy.  

Andrew: And nowhere in the training did they mention how actually to go to file > open. 

Beverly: Whoopsies.  

Andrew: It was just already open? So, I quite literally did not know how to open the tool. And so, show up, it was back in the days pre-COVID when travel was very serious, and everyone was doing it. And I was entering a project in its war room, red, triple red phase. So, we were sitting in a big war room, our project team, with the CIO funding that project. He sat directly next to me.  

Beverly: Oh, boy.  

Andrew: And I was introduced as the guy who was going to get development back on track, you know, et cetera. 

Beverly: The superhero.  

Andrew: The superhero. I did not know how to start the tool that I was supposed to work in with the CIO, who was my six-times-up boss. 

Beverly: No stress at all.  

Andrew: No stress at all. It went fine. I used the restroom a lot the first couple of weeks, made a call, Googled some stuff on my phone, and we were good to go. But that’s how I entered the industry. It’s been a fun ride since then. 

You know, a lot of hype cycles were navigated. Hopefully, a lot of real work was done, and I’m in the middle of that. As you indicated, I run the CPG and retail practice at Wavicle Data Solutions. 

I’ve been focused on that space for the last seven or eight years and smattered around a little bit of healthcare and a little bit of financial services. But I’ve been focused, like I said, in that space for eight or so years. 

I also work across the data space, from a services point of view: strategy, advisory, design, build, implementation, production, support, and as a person you can call to yell at when you get angry.  

Beverly: Yes, the throat to choke.  

Andrew: Right, exactly. All those roles. In terms of capabilities, at Wavicle and within my practice, some are related to those service offerings. But of course, we’re focused on some of the more traditional, foundational data management type practices, governance, and that type of stuff which can be a big accelerator to some of the cooler things that are evolving today.  

Generative AI being one, data science in general, advanced analytics, traditional BI, and enterprise scale, can be pretty cool. It’s difficult to run a business if you can’t understand your sales by state, as an example of some level of fidelity. 

And it’s been a fun time to be in the space. So, I would say retail, particularly, has been on an interesting journey through COVID and then coming out really front of line when you talk about supply chain, inventory, focus, buy online, pick up in store, and have to navigate to that and the acceleration and then add transitioning out of that. 

It’s been an interesting journey. I rode the wave. I can’t wait to write the book 10 or 15 years from now, but that’s what makes me cool.  

Beverly: That’s really a good reason to be cool. Those are awesome. And on top of all that, your education, I’ve heard that you’re a legal person?  

Andrew: Who told you that?  

Beverly: Is that true?  

Andrew: That is true. I am a reformed attorney. I practiced mostly in tech by trade as an undergraduate student and then had the delusion of grandeur that I would practice IP and law in that space. I mostly focused on text transfer but did practice for a few years. It was a little combination of things that kind of pushed me into consulting. 

One was, frankly, the job market around 2008, 2009, 2010. For those of you that remember back then, those were interesting times to navigate. And then secondly, and I mean this with complete love and adoration for my legal brethren, they are super important and needed to support industry, but I just wanted to work on the side generating the contract and generating the legal questions versus dealing with those legal questions. And I thought that was a pretty good indicator to look for a career switch. At that time, I basically was doing a little bit of boot camping on the side, SQL, Perl, and stuff like that, and I was given a shot at a little consultancy. And the rest is a minor, minor footnote to history. 

Beverly: I love it. It’s interesting to see all the backgrounds that we have in data science and AI. You never know where someone’s been and where they’re coming from. 

We’re talking about the top challenges and opportunities that gen AI brings to the retail experience. So, what are some first examples? Could this be something really simple when we talk about generative AI? Because when people think about that, they’re not quite sure what to think of right now. It’s like some giant thing, but it could be something really minor. Do you have a few examples?  

Andrew: Yeah, there are a couple of really practical examples. I think when folks think of retail, it probably means different things to different folks. Obviously, there’s a lot going on in that sector about buying clothes, buying groceries. That definitely falls under the umbrella, and that’s something we all deal with in our day-to-day lives. One space in which generative AI has been very helpful in both of those subsectors, and I’ll take grocery first, is offloading some of the thoughts around grocery list generation. What am I going to make for dinner on Thursday? Just a really great helper. 

I mean, it’s human-in-the-loop type stuff, but so many grocers, although they’ve learned from each other over the last 18 months, there probably is a use case out there if folks can Google where one of those recipe chatbots did start recommending arsenic after about six or eight months.  

Beverly: Whoopsies.  

Andrew: Got our hands around that. Don’t do that, kids, at home. But overall, interacting with the recipe chatbot, “Hey, eight mouths to feed, three kids, one dairy allergy, what do I make for a quick dinner on Thursday?” Not only does it suggest three or four things, but once you click, it generates your grocery list and can then interact with the fulfillment side of the house, pull those groceries off the shelf, and do a BOPUS right to the store. So, it used to take a lot of thought capital, especially after a long day of work. 

What am I going to make for dinner? I’ve got to get this spice. What is saffron? Is it golden? Whatever. That type of thing. And even budget constraints. So, it’s made some of those things really practical. It’s a lot easier to pick a recipe, try something new, generate a grocery list, and go. And kind of the analog over in the retail space. 

Obviously, there are a lot of chatbots and other interactive things that can suggest shirts, pants, shoe combinations, and things of that nature, if-then type stuff. But also on the customer service front, particularly when looking for inventory and size up and how to navigate those types of things, chatbots have made that a lot easier. So, if you’ve been to Zappos or Amazon, or particularly Walmart, it has launched a generative AI e-commerce chat assistant that they’ve been very public about, that’s generated a lot of conversions and a lot of traffic in all those places. 

Beverly: Nice. So, something as simple and common as a grocery store, you can use generative AI to help create a list and come up with a meal plan for the week.  

Andrew: Exactly. 

Beverly: It is easier than debating, arguing, and finding this item. Then, you also gave an example of clothing. Do you remember Garanimals? You’re probably too young for that. 

Andrew: No, I don’t think so.  

Beverly: Okay. Well, the guys used to use these. 

Andrew: That sounded like a great punk band. 

Beverly: Yeah. It was cute. They used to have mostly men’s clothing, but they would have a tiger on the shirt, and then there was a tiger on the pants. And so, you match up the tiger, and you would know that that goes together.  

Andrew: And you’d say, there goes a Garanimal. 

Beverly: That’s right. Exactly. In some ways, it sounds like generative AI can help with even clothing purchases and make things easier for people. 

Andrew: Absolutely, especially if you’re going brand-first e-com, which a lot of people do. So, North Face is serviced if you want to do BOPUS or if you want to get your pickup online. You’re in Chicago, as an example, where recording can be very helpful to coordinate an outfit and tell you where you’re going, or you’re going to Nordstrom’s, et cetera, et cetera. 

Beverly: Yeah. What about in-store or immediate point-of-sale, generative AI examples that you can give?  

Andrew: Yeah. Well, I think that’s emerging a little bit more, but there definitely is traction there, and also some medium-term future use cases that are generating a lot of excitement in the space. When you think about it, this kind of ties into a broader use case, generative AI for micro-customer segmentation and experience and journey enhancement. So, when you think about retail, you’re talking B2C-type stuff for the most part. 

So, there’s an opportunity to be very personalized. And, you know, the more personalized, the better the conversion, is what the data has shown over years. People want to see what they want to see. They want to buy what they want to buy. That’s the holy grail. So that’s happening digitally. Outside of this brick-and-mortar store channel, when you think about email campaigns, your e-com experience, that type of stuff.  

But in-store, tying into retail media, some of the content that you might be served up in the aisle or at checkout, based on your loyalty number or even perhaps what your basket analysis has been via OCR or other types of quick visual scan stuff serving up particularly relevant media content or experience. “Hey, did you check out the ham in aisle 16? That complements the cheddar cheese that you purchased.” You know quite well that type of thing. 

So, I’d say overall enhancing customer experience in-store. There’s a lot, and I think there’s another couple of major use cases. I’ll mention one that jumped to mind: inventory replenishment in-store. 

So, when you think about monitoring, video, and some of those things, we’ve experimented with shelf analytics for years, demand forecasting, and frankly, even sensors on shelves. But now we can, with an investment, really detect, “Hey, you’re in real-time, low on white cheddar rice cakes.” Just to pick a favorite example from the Simmons household. And you can do in-store replenishment and probably circumvent, in most cases, trying to flag down the Target employee in the red shirt that you can never find; God bless your Target when you’re looking to see if something’s in the back. 

Beverly: Yeah. Unless you pick on some poor schmuck that showed up in a red shirt and khaki pants. Yeah, that’s difficult. 

Andrew: That’s been me a couple of times.  

Beverly: Me too. I admit that. So, it sounds like a lot of it has to do with the consumer and centering around them. In the past couple of decades, we’ve moved from mass marketing to one-size-fits-all. And then we moved into segmentation, and then we moved into the hyper, and now we’re almost extremely personalized down to the individual and even maybe the individual in a certain circumstance, and the AI and generative AI are able to make those types of granular recommendations around the consumer. Sounds great. What could possibly go wrong?  

Andrew: Well, you know, a lot, of course. I mentioned the arsenic example, obviously, when it comes to content, especially when it comes to brands’ relationship with their customer base, which is the most precious commodity that they’re retailing. 

A brand is hard to build, and its relationship with its customers is also very important. So, there’s a reluctance to take a human out of the loop when it comes to the experience that you’re putting in front of the customer. So, obviously, as those bots start to drift and hallucinate, which will happen without proper attention and fundamental investment up front, from a maintenance point of view, that runs a pretty high risk. 

Overall, most retailers are taking advantage of use cases internally and trying to build up their competency. It’s an interesting hype cycle in that there’s definitely executive-level enthusiasm and excitement. At the same time, I would say there’s SVP-level and some director-level reticence because of some of those challenges. 

Ultimately, what I’ve been telling my customers is that if you think about your YouTube algorithm or your Instagram search algorithm, which for a lot of people, maybe me, is precious to them, if you throw something into the algorithm, that’s going to mess up what’s served. And you find very interesting things about what’s served up to you. 

I think Instagram is a great example of targeted ads based on content, at least for me anecdotally. I mean, it’s always the shoe, hat, or vacation experience.  

Beverly: So don’t let your spouse borrow your phone for a minute and do a search. 

Andrew: Which is another good reason. Yeah, I got to do that. And so, if you extrapolate that out and run it out 10 or 15 years, our view is that living life algorithmically can be a real benefit to folks. So, when it comes to whatever you’re purchasing in a retail context – food, clothes, convenience stores, home goods, whatever the case may be – try to push that same algorithm that you’re experiencing digitally into real life, but it’s happening digitally in the backend. The clothes that you like, and the shoes that you like. And I think we’re in an interesting push-pull moment here in the short and medium term because folks, to some degree, and I think they probably always will, enjoy that tactile interaction with what they’re purchasing and that human-to-human touch. And that’s great. However, our view is that it may start becoming boutique and niche. 

And we free ourselves up to be focused on more high-value tasks, spending time with family and whatever it is that we want to do, and a little less time doing the things that we don’t want to do. Grocery shop, pick out which shoe, especially. And even if you like shoes and you like spending that time, maybe you find value in finding just the right one, or the shoes that you never knew that you would like, or that type of thing. So, it’s an exciting frontier. And I think that’s where the weathervane is pointing.  

Beverly: Okay. So, to summarize, I think I can try to bullet out some of what you just said.  

Andrew: Yeah, please.  

Beverly: You mentioned the damage and the difficulties associated with losing brand equity or modifying the image that a company is giving when you’re using generative AI. You hinted about the emotional and psychological difference for a consumer who doesn’t get the satisfaction of doing certain things that they maybe didn’t realize they enjoy. 

Andrew: Absolutely.  

Beverly: I get to check a box because I found the saffron to give an example you gave earlier. And just the difference of going about life that way. 

So, if those are some of the most common damages, is the juice worth the squeeze? Like, is it with some people, retailers in particular, because retailers are sort of on the crest? Would they say, “Well, it’s just too dangerous. It’s not worth it.” Or are the benefits so great that everybody’s going to give it at least a shot?  

Andrew: The short answer is yes. The benefits are so great that people are going to give it a shot. I do happen to think that most of those use cases are somewhere between incremental and really positive growth as opposed to exponential enhancers for the business. 

There are a lot of internal use cases. Retailers, and this is cross-industry, retailers are the same as the financial sector and the insurance sector. A lot of internal productivity boosts – code generation, task automation, email generation, et cetera. But then, when you think on the customer side of it, it’s the same. You know, so you pick up an example from earlier. 

If you automate your Thursday night grocery shopping based on meals, there’s some incremental sales. You know, “Hey, would you like to try that golden saffron that you’ve never tried versus whatever saffron, pepper, or whatever we’re using?” But you’re probably not buying two meals or three meals because of that convenience. Those are what those use cases do. They help drive convenience, brand loyalty, and customer experience. But those are all incremental pieces. So, I do think brands need to think of it in that way. You know, it can be beneficial for the bottom line, grow the top-line margin, and grow your customer base. 

So, in that sense, if your competitors are doing it; you need to be interested in it. But you need to position it right from a forecasting point of view on your board as they read the New York Times and hear about this explosive thing called generative AI. It is interesting and game-changing in terms of how we live our lives. 

And I’d be interested to revisit this podcast in five to 10 years. There probably are some industries where it’s truly game changing. You see exponential growth or even new industries crop up, but real positive incremental uplift for, relatively, in terms of ROI, cost-neutral investment. 

Beverly: Okay. And so, if you think about if your project and then project some more, let’s say we’re 20 years out.

Andrew: That’s what we do on podcasts.

Beverly: That’s right. What do we see? What do we see happening? Like, is there just almost no human-to-human interaction anymore? Like are we with our retailers asking consumers to consistently, constantly self-serve? What’s that going to look like in the future? 

Andrew: Well, there is a lot more. I think it’s an open question, not to get too philosophical here on the TAG podcast, but I’ll even speak anecdotally for myself.

For some reason, when I interact with a chatbot, it feels fine until I know it’s a chatbot, and then it feels a little different. And I think, one, AI will get a little better at emulating a human. You may never know the voice is there. 

And I think, culturally, we’ll all get a little bit more comfortable with that. I believe that, according to personal philosophy, human connection is important. And so, I think there are probably some pieces that will remain forever. 

We’re in a little bit of an experimentation stage regarding what folks would prefer to automate and what folks would prefer to be face-to-face. Obviously, from the business point of view, they want to push cost takeout, so that’ll influence how that trend evolves over time. 

And people will get comfortable with things. But I do think there’ll probably be some guideposts that continue to be human-to-human interaction. I do think that there’ll be some that are not like going to a record store, perhaps today. 

It was interesting and novel for some folks, and it was fun as opposed to, you know, firing up Spotify just because that’s a fun interaction and a fun way to spend time.  

Beverly: I mean, I miss Blockbuster.  

Andrew: Yes, I was more of a Hollywood Video guy, but I get it. Overall, I’d say that’s a clean way to put a button on it. 

A lot of those loops when it comes to retail, you know, omnichannel thinking about buying, prepare to buy, buy, fulfill, review, buy again type of stuff. It may look like a digital music space, and most people are comfortable in that digital space. But there are still record stores, more of a boutique, more of an experience, but niche. 

Beverly: Right. Okay, great. And for our last question, what final piece of advice would you give to somebody in the retail space interested in implementing more generative AI solutions?  

Andrew: The number one thing would be to experiment internally first. I’d say that probably goes, well, certainly with data, but maybe even outside of the data space. Failing fast, but never failing in production, and certainly never failing in front of your customers. 

I think there’s a tremendous amount of benefit to be gained from generative AI, even within your IT shops. When you think about things like code generation and QA remediation, things that can be an uplift for an IT shop where you can start building up the competencies and demonstrate some internal wins are that type of thing.  

The other bit of advice I would say is to definitely fly cautiously. Still, for those folks who have navigated hype cycles before, there are opportunities to hitch your wagon to executive hype cycles and societal hype cycles to enact your agenda. 

Generative AI is definitely one such case right now. So, be smart and definitely set realistic expectations, but this is a great time to get, as an example, some foundational work done that’s perhaps always been on the agenda. If you can connect that to a generative AI use case that’s of interest to the customer base, as well as to your executive suite.

Hey, to get this, we have to do that. Right. And to folks that are managing a data analytics budget, be aware that they make some hay while the sun is up. 

Beverly: Right. Gotcha. Okay. Well, thank you so much to Andrew Simmons from Wavicle Data Solutions, which happens to be our sponsor of TAG Data Talk this year, for talking to us about the top challenges and opportunities generative AI brings to the retail experience.  

Andrew: Thank you for having me. It was a blast.

As generative AI continues to evolve, it’s clear that its potential to transform the retail experience is immense. While experimentation and internal innovation are key to success, businesses must also tread carefully to safeguard brand equity and customer trust. The future of retail lies at the intersection of convenience, personalization, and thoughtful human oversight. As this technology progresses, it’s not just about adopting AI but understanding how to balance automation with meaningful human connection. 

Explore the full catalog of TAG Data Talk conversations here: TAG Data Talk with Dr. Beverly Wright – TAG Online.