Shelley E. Kohan (00:01.776)
Hi everybody and thanks for joining our weekly podcast. I’m Shelly Cowhan and I’m excited to welcome Shiv Gupta today, who is the managing partner at Volte, Volut, did I pronounce that correctly? Volut Group. And also you’re the founder of Quantum Site, so you do a lot of consultancy in the terms of data and analytics.

Shiv Gupta (00:13.528)
Volute, volute.

Shelley E. Kohan (00:22.03)
and I know your core area of expertise is helping Fortune five hundred companies really understand data analytics and also how to drive customer loyalty through that. And another fun fact about you is that you you’re not just retail, you’re retail, healthcare and financial. So Shiv, welcome so much.

Shiv Gupta (00:42.488)
Thank you, thank you, and it’s a pleasure to be here.

Shelley E. Kohan (00:45.008)
It’s great to have you here. I think the best part about the work that you’re doing and why I think it would be so interesting for our listeners is because you’re really kind of bridging the gap between data, technology, AI, and the business decisions that actually drive consumer behavior, loyalty, and growth. So I know our listeners will learn a lot today. And before we jump in, it would be great if you could give us like the 30-second version of how you ended up at the intersection of AI and retail.

Shiv Gupta (01:15.756)
Yeah, well, you know, it is a thirty second story because AI is just showing up as a relevant with high relevance with within within retail. and it really started with a personal story. a few months back I was about beginning of the year, I was hunting for a travel card and a new travel card, new travel credit card with benefits and and all that. and I went to my banker who I’ve had a relationship with for multiple years.

And I asked her, well, what do you guys have? And she gave me her premiere card, gave me the pitch. and this is the bank I’d had relationships with for better part of a decade. but it just didn’t sit right. I wasn’t ready to pull the trigger on that card. It came with a high fee and so forth. So I asked the AI agent, you know, what do you think? Is this is this worth me getting into? Are there alternatives? And within one minute, it was able to identify a much better card for me.

It even told me exactly why I don’t want the other card, my travel habits, the way I redeem. And so I think we all have to sit with that example, and we’re all gonna have those examples where you have a relationship over 10 years, a personal relationship with a banker over multiple years, and all of that was interceded in one minute by an AI that understood me better and actually directed me to a card that actually made more sense for me.

Shelley E. Kohan (02:17.316)
Crazy.

Shiv Gupta (02:39.916)
That was very powerful. And that’s when that moment hit me. I said, no, we we’ve got to deal with this as as marketers. this is powerful, what just happened to me. and that’s exactly how that that you know that that journey began for me.

Shelley E. Kohan (02:54.424)
It’s very interesting because as we’re hearing, and and you hear stories like yours, Shiv, which is a great example, that these agents are really helping to pre-select, find, give advice on shopping on in terms of the retail sector. They’re doing a lot of other things outside of that. But so now we have this new shopper who doesn’t really trust websites and doesn’t really kind of collects data from many different sources. So let’s talk about.

how brands are going to optimize this.

Shiv Gupta (03:28.792)
Yeah, absolutely. And I think brands have to think about this from a a branding perspective of the human branding and the AI branding. And AI is probabilistic. And I think if we if we focus in on that question that’s actually looking to answer every time there’s an interaction with a potential customer, it’s saying what is the probability that this brand, this product will satisfy the needs of this customer? It is a probabilistic model. Now, if you if you step back from that.

understanding and say, well, how do I drive a probabilistic model? You want to provide data, facts, information that confirm your ability to meet that need because the AI agent is saying who has the highest probability or what maybe two or three products have the highest probability of meeting this need. And you want to be able to feed that. So branding isn’t emotional when it comes to an AI. It is much more factual, but the emotion doesn’t disappear.

So the other part of it, people say, well, is it completely devoid of emotion? No, it’s not, because somebody will say, I want a brand that makes me feel like it’s safe. I want to, you know, there’ll be there’ll be expressions from the customer that will describe emotional content. Now, even there, the AI agent is taking a probabilistic model and says, what is the what are some of the factors that I see in the information I can gather that describe safety, that describe comfort, that describe trust.

and it will actually try and quantify that. So it’s not devoid of emotion, but it’s taking emotion and putting it into an algorithm.

Shelley E. Kohan (05:05.21)
So and the way that the agent works is so they’re I know they’re collecting a lot of structured data. Are they also collecting unstructured data? And how are they getting that data, especially from brands?

Shiv Gupta (05:17.228)
Yeah. So there is you know, there’s a whole space of AEO, S and GEO, which is all about commuting to the communicating to the AI. and so that’s that’s an important step. You can’t, you know, you can’t avoid that. But what they’re also gathering is a lot of information on your unearned media. So this is review sites, articles. Sometimes it’s going to be you know, an influencer who talks about this and and and does a blog post on it. There is information that it’s gathering.

fairly broad, that is going to also influence the decision. So it’s not just about what you say about your brand and your product, but it’s also what it’s hearing from the larger ecosphere about your brand and your product. And it’s reconciling that. And in that reconciliation is where that decision really gets made as to whether that product gets elevated in a recommendation or completely ignored in that recommendation.

Shelley E. Kohan (06:13.84)
So I know one of the big issues is that we have the agent-referred shoppers who typically have a high intent. They’re absolutely looking for something. So the intent is there. But as you say, the current infrastructure fails to capture them. Tell us about that and why that is the case today.

Shiv Gupta (06:34.702)
So partly I think we’re still working through how to describe our products to the AI agents and our services, right? So that’s still being worked on. There is a belief, which is the incorrect belief, that if you speak louder, you’ll be heard. And that’s just not the case. You need to speak smarter. obviously you need a media presence, but being louder doesn’t make a huge difference for AI agents. Second part of it is I think AI agents are still not able to.

understand the full chain of activity that’s required. So it may recommend a product. That product is out of stock. Or that product then can’t be delivered for three weeks or four weeks, or it didn’t account for the fact that you have high delivery charges, right? So those types of things still get in the way. I’ll give you another example. I was looking for a polo shirt and I said I don’t want to spend too much. Give me a polo shirt in the $50 range. And by the way, I’ve as I get older I’m kind of

over the synthetic material, I want cotton, I want to kind of breathe. and and I I count that as my luxury natural fibers. So I said I I’d like natural fibers. And I usually end up going to a a major retailer that you know we’ve all heard of. I’ll I’ll withhold names to protect the innocent, mostly me. But

Shelley E. Kohan (07:51.067)
Sure. There’s only one really famous polo polo shirt brand out there, but we won’t say it. Yes.

Shiv Gupta (07:57.667)
Well, i i I yeah, but there’s several that provide decent ones, right? and it it it steered me away from there. And it said, I know what you’re looking for. They don’t always lean in on natural fibers, you may find it there. They’re not gonna always have your price range that you’re looking for. They do sometimes on sales and so forth. It said you’re gonna have to work harder if you go to that retailer. Here are two other retailers that will fit your requirement. Now, what was interesting is one of the retailers that it recommended.

Shelley E. Kohan (08:04.684)
Interesting.

Shiv Gupta (08:27.586)
Had a very more cumbersome return policy. And so when you try something from an online retailer, you want to make sure that return policy is easy, right? The other retailer was not anywhere near me. So it hadn’t accounted for that, right? So there is there is still a lot of information that’s missing that helps AI agents make the best decision, but it’s still a very powerful influencer. And the more I think retailers, service providers start to plug in that information.

Shelley E. Kohan (08:31.594)
Shelley E. Kohan (08:44.804)
Yeah.

Shiv Gupta (08:57.358)
So that it’s easy to access and understand, then you’re going to get a lot better recommendations that come out of that. Part of this is also understanding a whole new set of metrics. The metrics that we thought about traditionally are not going to be relevant. You have what I call trust metrics. And these trust metrics aren’t simply saying you can trust me or I’m rated well, but you can trust me to deliver on that particular problem.

Shelley E. Kohan (09:12.28)
Absolutely.

Shiv Gupta (09:27.308)
You can trust me to deliver on this particular segment’s needs. and so you need to start to think about what are the trust metrics that I you know, and they are KPIs. They are the new label of KPIs for agentic buying. what are the trust metrics that I want to see that I want to convey to the AI universe and how will I manage them internally? How do I track them internally? How do we know that we’re generating those trust metrics on a regular basis? So that that’s another important part of.

of getting to that. So it’s I know it’s a long winded answer to your question, but I think it’s it’s it’s really the the complete story.

Shelley E. Kohan (10:03.512)
No, it makes perfect sense. So tell us, so we learned actually a lot. I don’t I don’t think we intentionally set out to learn, but Black Friday last year. And you know, a lot of retailers this year, we’re already gearing up for holiday selling. It started, you know, in I don’t know, June this year. Keeps going back and back. Prime Day started so much earlier. But let’s talk a little bit about what was learned from Black Friday from AI shopping perspective. What what happened then and what can we kind of move forward?

with understanding going into this holiday.

Shiv Gupta (10:36.878)
Yeah, I I think you one, we we learned people want help when they start to shop for for their relatives, right? So that that basic aha was like, clearly the market hadn’t been meeting the need, right? We we’d seen a lot of a you have AI agents on behalf of retailers and they try and make recommendations. What I think

We’re gonna learn, and the aha is going to be is that those agents are limited. They’re limited by the interaction that that individual has with that platform. Where whereas your AI agent, whether it’s ChatGPT, Anthropic, whatever you’re using, that has a history of understanding you. Potentially even personal information about you and your family and your interactions with your family. People will even confess, boy, I need to buy something for my sister, my brother, my cousin, but I don’t like them.

and I don’t want to spend over 20 bucks for it, but I gotta do this. Right? They’ll actually confess to the AI in that situation where they won’t confess to Alexa or or someone else that’s that’s on that platform. So you have a within even AI, where retailers are so focused on implementing AI for their platform and and to be able to communicate to their customers, I think they need to recognize there’s a differential advantage to the AI that the customer is using.

Shelley E. Kohan (11:36.56)
That’s so true.

Shiv Gupta (11:56.717)
Because it has a deeper history. And that deeper history will inf provide better information. So that’s, I think, the other next learning we’re gonna we’re gonna start to recognize is that people went to their AI versus the platform’s AI because they felt like their AI had a more holistic picture of their needs and wants and could contextualize information a lot better than than the platform’s AI.

Shelley E. Kohan (12:24.048)
So what should retailers and brands do knowing what you just described going into the holiday season? Like it sounds almost as if they have no control over this, but do they?

Shiv Gupta (12:36.022)
they absolutely do. So here’s a few things that you can do. And I almost think back to was it the Jay Peterman catalog back in the Seinfeld days? Yes. There is this narrative that points, you know, to a a romanticized use of the product, which I thought was interesting. Now I don’t think you can take that immediately, but it’s it’s part of that, but also painting a story on the product. Like where does it w who would like it? How will it get used? What is the

Shelley E. Kohan (12:44.665)
yes. I love that. So funny. So funny.

Shiv Gupta (13:05.218)
you know, what is the the functionality of the product or the service? Those types of things I think will be much more important. So labeling your products, especially the ones you think are very good or are important, or your your leading product, you definitely definitely want to start to define the space for that product with as much clarity as possible. so that’s part of the AEO GEO side of it. The other part of it though, and this is where

It’s not thinking about just Black Friday, but it’s thinking about the next Black Friday, then next Black Friday. This is your opportunity to set up a long-term pathway either to success or potentially a deep well that you need to dig out of. And what I mean by that is let’s assume the AI product makes that recommendation. You’re going to get feedback. And the more the AI product is recommending, I mean the AI solution is recommending your product, the more likely it is going to get purchased.

The more experienced data is going to be generated around it, and the more reviews are going to come back. And if your full execution isn’t well done, it’s going to get noticed. And at that point, future iterations of this are going to analyze that information and say, boy, I made this recommendation for this product, but everybody seems to say it doesn’t live up to that. It breaks too quickly. Or, you know, this has a lot of some they lied about the ingredients.

Shelley E. Kohan (14:24.484)
Mm-hmm.

Shiv Gupta (14:30.198)
On that list, right? Anything of that nature. Now you’ve got a trust issue. And these are the trust metrics that AI has. And AI is again looking at its probabilistic model. And if you can even imagine, as a human being, you’re doing not a mathematical probability, but just sort of that guttural probability estimate. If somebody’s failed you once, you automatically know that the probability of you trusting them again is diminished dramatically. And the same things happen with AI, because AI has

Tons and tons of choices, it doesn’t need to give you the benefit of the doubt. You have to earn the benefit of the doubt out of it. You know, so any hole you dig yourself into, you’ll have to earn out of it. So think a little bit, not just about how do I get this sold today, but how do I make sure that the feedback that will come out of this event will be positive, and therefore I’ll be well positioned for that next round of recommendations.

Shelley E. Kohan (15:23.482)
Shiv, it’s like they have to play offense and defense at the same time.

Shiv Gupta (15:27.232)
It’s getting to be a much more complex world. And I think the insight I have for retail is I think niche retailers are going to do a whole lot better than they’ve ever done before. because they are able to define the product, the audience, the service with a lot more clarity because they don’t have that spread of assortment. I think from that perspective, AI is is definitely a gentic buying is definitely a friend of the niche retailer.

Shelley E. Kohan (15:31.376)
Mm-hmm.

Shelley E. Kohan (15:40.496)
Yeah.

Shelley E. Kohan (15:54.329)
I love shit. First of all, thank you for throwing me back to Jay Peterman catalog, because I love that. And then Elaine went to work for Jay Peterman. And that was like just hysterical. From a retail perspective, I really enjoyed that part of Seinfeld. But here’s a scary number: 80% of consumers are gonna use Gen AI to shop with this year. 80%. That’s tremendous.

Shiv Gupta (16:01.816)
Yeah.

Shiv Gupta (16:14.531)
Yeah.

It it it is tremendous and it you know, that’s great retail statistic. We’re seeing the same risk retail I mean similar statistics in financial services. We’re seeing similar statistics in healthcare, shopping for a doctor or a dentist, or you know, a procedure. all of that. This is this is fundamentally and this is one of the things I I I tell my clients is this is not search two point What you’ve entered is a fundamentally different.

Shelley E. Kohan (16:43.15)
Right.

Shiv Gupta (16:46.09)
era where you have an intelligent intercedent who’s going to help make decisions and you may or may not be on that list and no and your amount of advertisement, your amount of brand presence is not relevant. One of the other things I found that was interesting, I you know checked when I started down this path, I said, well, celebrity endorsements, do you care if a celebrity’s endorsed a product or not? And the clear answer was no, I don’t care. It is not a factor in my decision making process.

Shelley E. Kohan (17:15.245)
Interesting.

Shiv Gupta (17:16.536)
Yeah. So a a lot of that major spend that we do to get that exposure because the celebrities now endorse that product, you’re really relying on that human chain to carry you through because the AI chain will not. you know, unless you can make that celebrity experience highly, highly relevant to convey a specific set of information. So for example, you have a famous golfer that says, Look, this is an amazingly good club, and this club

Shelley E. Kohan (17:29.455)
Yeah.

Shiv Gupta (17:42.713)
You know, has these features that make it work for me. And when I coach others potentially, I let them know based on their size and their gait that this is the right club for them. Like that information now from an expert makes sense. But if it’s simply I, you know, I trust brand XYZ, you’re you’re not going to get that return. and and I think that’s the big shift in branding that needs to happen and happen.

Shelley E. Kohan (18:08.144)
Yeah.

So, Shib, I wanna go back to something you said earlier. You said something about how consumers will kind of confess or say things to an agent that they wouldn’t probably say face to face. And I think there is, I’ve always said that in retail, as long as I’ve been in retail, that what the consumers say and what they do are often very different. And I don’t think it’s like purposeful, you know, trying to act differently than what they say, but I genuinely think.

They they wanna they wanna do this, but they end up doing that. So tell us a little bit about the data gap between what your customers are saying but what they’re actually doing.

Shiv Gupta (18:48.834)
Yes, you know, I I think that’s a very important gap. And that gap’s existed before AI. So we can even move on from AI here. That’s just a very important gap. You know, I I think part of that has to do with our market research methodologies. We ask based on, we ask questions or we launch surveys to try and get those consumer insights based on our ambitions for our brand, based on our desires to see where the brand can go. But they’re not, they’re always prompted in some way or the other. They’re not unprompted.

I think there’s a middle ground between the behavioral data, which I think is super important because it really tells you the truth about you know how many people will consider a product, but then how many end up actually purchasing it, right? but other than that, there’s also unprompted data. And I think that’s the bridge. It’s understanding what people say about your product when you haven’t asked them a specific question. And

Right now that is available on review sites or other sites where you don’t own that property, but even in owned properties, I think that that opportunity is is increasing. because of AI and the way AI chatbots can be used by retailers is to get information about why somebody is there, their context, which is often missing. Because what the stated behavior is and what the actual behavior is is driven also by context. And the context will be pretty important.

Shelley E. Kohan (20:04.154)
Right.

Shiv Gupta (20:11.272)
one of the things we did a while back was we were working with a retailer of sports equ you know, sports machine machinery, so like small mobiles, things like that. just sort of fun activities, but and and so forth. So it was interesting to find out the brand itself felt like their big exceptional play in the market was to talk about innovation. They’re an innovator.

Shelley E. Kohan (20:20.794)
Mm-hmm.

Shiv Gupta (20:36.6)
But when you saw the unprompted responses from their customers, it was all about fun, excitement, having fun with family, being outdoors. It was all about the experience and what it enabled them to do. And friends and family played a big role in that. Not one, not one person mentioned innovation. It just didn’t register.

Shelley E. Kohan (20:56.656)
Well it it’s so interesting shiv because I don’t think the consumer says, I want to go shop at an innovative company. You know? It’s not their language, right? Yeah.

Shiv Gupta (21:05.422)
Exactly, exactly. But but the company does. The company thinks I want to be known for innovation. Well, nobody’s gonna care. Like when you really look at the unprompted data, your message on innovation is not gonna matter. what’s gonna matter is how do they experience that innovation, and can you start to speak to that? That’s the value of that unprompted data. Because I can guarantee you, if we’d launched a research project that said, Let’s talk about the importance of innovation, everybody who responds will say, Of course innovation’s important. I think of innovation as

Super important. I want an innovative company. I like new features. They’ll tell you all of that, and you will falsely believe that that’s an important aspect of what you’re delivering. And it may be behind the scenes, it may be within your company very important, but it may not be important to how the consumer perceives your products and services. And so I think that part of it, the unprompted data, is a rich land, you know, it’s it’s a rich mining opportunity.

Shelley E. Kohan (21:53.199)
Yeah.

Shiv Gupta (22:03.48)
to get much more insight and make the connection between behavioral data and stated intention.

Shelley E. Kohan (22:09.594)
Yeah, I think that’s great. So the other thing I know you always profess is that, you know, being data rich doesn’t mean you’re insight rich. And I love that because, you know, you can have all the data you want unless you’re doing something with it, right? And one of the things that kind of shocked me, one of the statistics that really surprised me, I’ve been teaching customer lifetime value for, you know, three decades. It’s a basic metric. Yet I read recently that less than 50%, a little over 40% actually, so way less than 50%.

of companies can actually accurately measure customer lifetime value.

Shiv Gupta (22:46.784)
It it it i it is it’s more complex than it seems, honestly. I think that’s that’s the the problem. Like when companies decide to focus on lifetime value, they don’t realize how multidimensional it really is. first lifetime value could be, how much has this customer spent in this organ? So like an RFM history. But that’s historical, that’s not prospective.

So you also have prospective lifetime value. Can you model when you meet a new customer? They walk in the door. Can you estimate their lifetime value today? Because that would define how you treat them, what services you can offer, what they might be interested in. So that’s just going from past to future looking LTV, that most companies don’t do. The other part most companies don’t do is they don’t look at a time-dependent lifetime value.

Meaning if you are a new customer or you’ve been here 10 years or five years or three years with me, you may have the same buying patterns, you may have the same demographics, but nonetheless, your tenure with me also defines how likely you are to be lifetime, you know, what your lifetime value is likely to be going forward, and how much trust and equity I’ve built in this relationship with you. So a time-dependent lifetime value is another dimension that that often gets missed.

And I think when you start to start to refine lifetime value from these dimensions, you start to understand that there’s a lot more ability to differentiate and start to treat customers in a more trust, I’d say a trust aligned methodology than what a simple LTV number actually does. but it’s complex, right? You have to build your business around the idea of

I generate trust and that trust is equity. And then that equity needs to be then, you know, the formation of the value exchange I have with a with my customer. and so, you know, it’s it’s not just simply one metric, because if it’s just how much have you spent with me or an RFM analysis, that that that that doesn’t stick, right? That’s just very narrow view of LTV.

Shelley E. Kohan (25:00.314)
Yeah.

I like what you said about today’s customer and tomorrow’s customer, because the c customer you have today, this consumer that you have today, they’re they’re changing also. And so as they go through the next three years, that same customer is kind of a different customer. So if you’re not kind of forward looking in how you’re projecting out that customer lifetime value and those shopping behavior changes and the shifts and all of that.

Shiv Gupta (25:12.536)
Yes.

Shelley E. Kohan (25:26.024)
that could really lead to a lot of misunderstanding about your consumers for sure. So the last thing I want to ask you, and this is super important for our audience, is is that what so you do a lot of work with other industries, and I feel like sometimes in retail, we’re very siloed, we kind of stick to retail. There’s a couple retailers that are really great about looking at other industries, and those are and then they bring stuff in. So what can we learn from other industries? What advice would you give?

retailers about what we should be thinking about in terms of things that are happening in in other industries.

Shiv Gupta (26:02.594)
Yeah, I think one of the big trends I’d say in financial services, which I think are probably highly applicable to retail, is the idea of simplifying product and assortment. so it’s it’s it’s the more complicated a a buying decision is, the less likely you are to convert. And financial services understands this. I’d say the other thing from financial indust from financial services that I think is is very important.

Well, yeah. So we’ll stick to financial services and I’ll talk about healthcare in some ways. is is the idea that trust is an asset, and that asset then needs to be invested in in multiple ways. I think retailers don’t fully grasp that all the time because they they see things as very transactional. You need, you know, I’m fulfilling the need for that moment. And that’s that’s the nature, right? That’s the nature of retailers.

Ne retail needs to deliver a product or service that solves that problem that’s that’s acute, that’s today. Whereas financial services has the ability to build out more relationships. But I think in that near-term focus of converting, retail misses the opportunity oftentimes to think about building trust that is repetitive, that that can lead to a greater sense of loyalty that doesn’t involve points and discounts, right?

And so that’s part of it. And then with healthcare, one of the things I’d say that I think is important is healthcare is starting to understand the need to have advice on complicated you know, in their case, complicated purchases. right, so any sort of a knee surgery, there’s a lot of consultative process that goes into it. And to defer your consultative process to just your chat bot.

Shelley E. Kohan (27:40.74)
Definitely.

Shiv Gupta (27:52.832)
I a for a retailer, I think, is is problematic. It’s it’s not sufficient. or the chatbot needs to have enough information. So it it kind of comes back to that full circle back to what we were talking about with AI agents. You need to give AI agents the ammunition they need to be able to have that conversation in an intelligent way that’s highly personalized. and oftentimes that information is missing.

Shelley E. Kohan (28:16.654)
Thank you, Shib. It has been great having you on the podcast. Do you have any closing like points you want to make before we leave our audience?

Shiv Gupta (28:24.94)
No, I I I’d say that you know, AI is in some ways a revolution, but it’s also an evolution when it comes to business, especially if you’re analytically inclined, right? We’ve seen we’ve seen this coming in many ways when we did multiple you know, machine learning and and and probabilistic models and predictive modeling. So this isn’t this isn’t new. So if you’re not into analytics, I’d say take a step back and recognize this is an evolution, not a revolution.

and that there’s still time to start to make those changes and keep up with what’s happening. It’s it’s not it’s not the world is is completely changed today, but it is moving rapidly. That evolution is quick. So it may feel like a revolution, but it’s still an evolution.

Shelley E. Kohan (29:09.36)
Well thank you so much and really appreciate you being here with us.

Shiv Gupta (29:13.113)
Thank you. Thank you for your time.