Shaping customer experiences with agentic AI & personalization

Steve Phillips [00:00:00]:

Welcome to Inside Insight, where marketing strategy meets consumer truth with your host, Steve Phillips. 71% of consumers want generative AI integrated into their shopping experience. So what's driving that demand for better personalization? It's not just about chasing novelty. Consumers are feeling pressure from rising costs and they want help making smarter choices. Commerce today isn't just about what people buy, it's about who or what guides them along the way. That's where agentic AI comes in. It can act like a personal shopper, understanding both preferences and price pressures. The goal: experiences that feel truly tailored and empathetic. For marketing and insight teams, there's a huge opportunity here. It's about reading the signals accurately and being ready the moment the customer journey starts.

I'm Steve Phillips. Joining me today is Michael Nevski, Director of Global Insights at Visa. Today we're going to discuss how transaction data through a lens of agentic AI can help brands respond with empathy, relevance and speed in a cost-conscious world. So, hi Michael, welcome to Inside Insights. I'm excited to have you on the show. Can you start by telling the listeners about yourself? And please tell the listeners more about what kind of data Visa collects and how do you build shopper profiles and partnerships with banks and retailers. Give us an overview of the company process there.

Michael Nevski [00:01:33]:

Steve, thank you so much. Thank you for having me. I'm excited to see you to be on your show. Thank you. First and foremost, let's talk about me and my background.

So I work in several industries in retail, syndicated data companies like Serkana, led marketing research in the wealth management companies like EBS Financial and Schwab. I also work in manufacturing and primarily, like I call myself, I'm a recovering marketer because at some point in my career I realize what you and me are doing. We're identifying strategic white places and acting, as I like to say, as a neck for the head to helping our brands to turn to the right direction and navigate and help them to achieve those results. So it's a very strategic position. Same time, I also done some shopper marketing consulting and product management.

So all of those factors helped me to lead strategic thought leadership, consumer insights research for the global economic team at Visa. So, and with that said, when I joined Visa almost five years ago, I worked directly with the chief economist and pretty much I love my lane, I swim in marketing research lane as yourself. So I love that aspect of it. But same time I work with lots of economists who look at lots of economic data. But pivoting further to your question of what kind of a data is available to me?

What kind of a data my company process? Primarily it's transactional data. As you can imagine, it's a payment network. But same time, the challenge is always that even if you partner with issuers or merchants, you don't have a demographic per se because it's a true network. Right.

It's not a credit card company like some of the competitors. So we don't issue cards, neither debit nor credit per se directly. So that means you need a constant department with third party providers or some other agencies in the industry to be able to identify those trends. And primarily when you rely on internal data. So that's the transactional data, you identify the patterns.

Right. You can do some geographies, you can do some categories of shopping, and then you start identifying those patterns in terms of how often those cards making purchases, what type of purchases, and then you go and partner with some other partners to really identify segments. Right. In terms of the demographic profiles, it would never be achievable. In terms of the granular level, it's never done at the individual level.

It's always depersonalized and aggregated because we don't have such data. Right. So that's why. And plus, my company is very conservative in terms of how we process and store the data. That's why it's always at a level of trend, identifying that trend among consumers.

In my work, I also rely on traditional marketing research, right? That means surveys or focus groups or ideas to really identify new trends, like for example, rise of a single person household or adoption of a generative AI or agentic AI. And that one comes directly from marketing research and market research efforts, as you know. So hopefully that describes my function and that describes what type of data, right.

So we can collect or analyze, right? Thank you. Yes. Yeah, yeah.

Steve Phillips [00:06:18]:

So you're getting, you're getting an awful lot of data, an awful lot of transactional data. You're tying that in through partners with two other data streams and then you're analyzing that data. So how do you use that data for macro level insights, spotting big consumer trends, micro level strategies and informing specific campaigns. How is that used on the ground by both yourselves and your partners? Right?

Michael Nevski [00:06:39]:

Absolutely. That's a great question. Although I'm not doing lots of direct work in that respect. But for example, cross border travel, it's a macro trend. Very often you look at, again, at the trend, not maybe segmentation because you don't have a demographic data, but you can see where, what corridors people traveling from Country A to Country B.

Right. So and then you try to understand what destinations are most popular, who is traveling, what nationality terms the countries. Right. So and then you can advise your merchant partners, right, on the trend of understanding how large is the group or person in the sense of like is it like a company of people who traveling or maybe individual, brand, bad people traveling. But it's always at the trend level.

So there is no granularity for privacy purposes for lack of such granular data, unless you partner directly with any particular merchant trying to identify those segments. Again, it's all segment level. But really that's one of the examples when you're helping identify those travel trends and what's going on. Maybe one country is pulling back for any particular reason, which sometimes it's a geopolitical, economic reasons and stuff. Or maybe more people traveling to those destinations because they feel it's more accessible.

You overlay that with regular research and trying to understand the semantics and stuff like that. That would be the example.

Steve Phillips [00:08:07]:

That sounds great. So in that example in the travel industry you're obviously collecting data in real time. You're getting incredible levels of understanding of what people are doing, where they're spending the money, who are the people you're working with? Is it the hotel chains, is it the airlines, is it the retailers in the destinations? How are you serving that data to people?

Michael Nevski [00:08:26]:

Yeah, absolutely. That's what Steve. Exactly.

That would be type of merchants like you just mentioned and pretty much you support them in terms of their campaigns or efforts. Right. If you know that a lot of Belgians traveling to Cancun, Mexico, for example. Right. So that means you provide more messaging.

That's not messaging but information sites. Right. On type of travelers. But again it's a very high level because. Let's take the analogy from.

Step back for a second. So for your audience to understand. In my previous work I managed the consumer panel for Circana Worldwide and when we talk about the CPG world, which you're very familiar with, it's a very granular level because you have a SKU level data. Right. You have individual SKU, you have a consumer panel, you know, very granular transactional data in terms of what category people buying. With us, it's a generic.

Right. So it's more at broader category of travel. You don't see any items, you don't know what people buying in the sense. So that's why it gives you more information in terms of the trend where they travel. Right.

You can understand where they stay again at the trend level. But to your point? Yes. If you know that there is a stream of tourists, travelers cross border to this destination from this country, so either airline industry or hospitality industry. Right.

They can get better understanding and then maybe diving deeper for their own research or teaming up with us. So to really understand why people traveling there, what is important to them. Right. Those are means done through regular research by asking those questions, by connecting with consumers and understanding their impetuses or desires or needs. So that's how usually that's the approach.

Steve Phillips [00:10:28]:

That's great. And I moved to extending the travel example. So there's been a lot of talk about the Chinese tourists in the UK, where I am, changing their travel patterns. So they used to always be on tour buses, now they're doing more solo traveling. And presumably you would understand those trends, you'd be able to see that type of data coming through how and where they're spending.

So how does that inform smart creative teams? What are the smart creative teams, are the people using the data, doing with that data to create campaigns, to create ideas, new innovations to utilize that data?

Michael Nevski [00:10:56]:

Right. Again, great question. While, uh, listen, while I'm not really directly involved into that process because there is a sister organization which supports the brand and the company internally inside Steam.

But same time again at the category level, if you know that people going to certain locations, the Downtown Abbey or some somewhere. Right. And you know those because you see that. Yeah, because. Because you see the spend.

Right. In terms of like trends where the transaction happens. Right. Or happening. Then again on a trend level, you can incorporate some messaging to those destinations, to those particular sightseeing sites.

Right. So and pretty much make it that more appealing if you know that there is from country A interested in seeing those particular places in the country B and that could be influencing the creative because you can incorporate the images, you can operate video materials and stuff like that of messaging and activation. So but that's how in general those companies would approach that.

Steve Phillips [00:12:04]:

So great. And moving away from the travel example. So presumably you're also seeing for instance within food and beverages, you're seeing. Do you get to see data about what flavors this SKU level is or are you just looking at the macro trend? So you're seeing what types of drinks people are buying more often or less often, what types of snacks people are using. And that data must be invaluable.

Michael Nevski [00:12:29]:

Yeah, but in that particular case, no, again it would be the macro trends.

Right. Companies like my former employer, they would see such data. Right. Whether they go to Tesco or in UK or Care for in France, right. And what kind of groceries they bu

SKU level, again, I wouldn't be able to see that. Right. So it's a very high level macro trend because you see what it's called MCC code, merchant transaction code, but that's about it. So that's why, plus you don't have a demographics, that's why you try to identify trend at kind of a transaction platform level. Right.

And you understand, for example, if this card resides in certain country but then it travels cross-border and spends in certain destinations, you can help identify those trends or again for your business partners and deliver value as such. But it wouldn't be the kind of a SKU level information. And as you know, like in payments, in my opinion, CPG retail data is highly concentrated as well. There are only few providers who have access point of sale SKU level data as you know, in North America, Europe, in Australasia, et cetera. So that's why, again, it's a focus more on identifying and again in my work it's mostly focused on identifying early on trends, right.

Within the consumer segment, whether it's North America or Europe or worldwide. Like I mentioned, those trends are most interesting because very often my business partners — because my work is outward, right. It's about the thought leadership. And very often as yourself as I appreciate you or CEO of Zapi now, you do more creative and innovative stuff for that. You know, you're very much focus on your business and what's going on within your business.

And when I come in and talk about something I've noticed recently because like now it's already on air, solo traveler, right? But year and a half, two years ago it was a still nascent kind of a segment and even the rise of a single person household, for example, over 30% of households in the US and Canada already single person households, if we're talking about percentage. So that kind of created this kind of a light bulb came on and many assist with executives I talked to saying, oh, we didn't pay attention to that because it changes how people consume, how they travel, what perks they need to offer. Just recently some of the cruise liners started waiving those single person cabin fees because they don't feel it was like in the news recently because again they don't fill the ships fully, right? So, cruises, and then they're saying, oh, I'd rather take a hundred single people, right,

But with that fee because they'll travel with me instead of just charging them extra because they don't have a second person. Right. But also like services tailored specifically for a single person traveling and sightseeing. And that's where I think the biggest value I can bring and my company identifying those trends early on in communicating those trends to our business partners. And, and then they're, they're using that to amend, update, innovate their product offerings, their services, messaging.

Right, branding. So all of those in the box. So you absolutely right. And creative. So definitely.

Steve Phillips [00:16:16]:

Yeah, that's great. Let's talk about something that we're both passionate about which is agentic AI. And we, we've talked about this before. Personal shopping agents. How do you see these shaping customer experiences over the next—I was going to say two to five years but no one can, no one can look. So let's say one to two years shaping experiences. And what do you think marketing leaders should be preparing for as this agentic shopping comes online?

Michael Nevski [00:16:44]:

This is a great, great question Steve, because I just finished a custom study in Canada on agentic AI and key insight. I would say we have a high hope that agentic AI will improve my life and help me.

And over 90%. Over 90%. 9 in 10 people saying I'm still fear of this stuff. It's unknown. I still don't understand that.

It's kind of one of those in the primary times when moment or something we see that, oh it's interesting. I probably can use it for food but same time I'm scared of that. Right, Something like that. Maybe not a good analogy, but yeah.

Steve Phillips [00:17:29]:

No, it's opportunity simultaneously.

Michael Nevski [00:17:31]:

Yes, exactly. We still, if you remember that Crossing the Chasm book, great book I read in MBA over a century ago. So that we're still at the very early innovator stage. Right. As we ramping up another one which is very interesting, I asked this question: would you willing to share your information with agentic AI or with the bot which is going to act on your behalf to procure products and services?

So in about 15% said yes, but primarily for specific task, because when the consumer is being asked to share their information in general like a feed, very low percentage say yes, 2, 3, 5%. But once it's a specific task, go and acquire this pair of jeans for me then I'll give you my data specifically for that task. But what's interesting, the next step further and it's kind of elasticity scale. I asked the question: if agentic AI will help you to realize 10% savings, would you be willing to share your information? The percentage almost double from 15 to 28% said yes I would.

If I can save some money, you see. So this incentives and of course it will hit some flutter eventually. Right. So but this is kind of a setting initial settings to your question what do you think how it's going to influence. I think once we threw that early adoption I think once the NSI my personal view on that, I think same like with generative AI right now we're all trying to adapt to it.

Some of us using it even at work. Right. So try to see if it's help us to be more efficient. I think once we pass that early innovators stage and move to early majority once we know that technology is reliable, technology is efficient and we know that there are settings where if, let's say, transaction was made in error, consumer knows that they're protected. Right.

And who's responsible for that? So consumers feel comfortable with certain transaction limits, financial limits. So then I think it's going to be much more ubiquitous for us, like in a sci-fi movie, to rely on agentic AI to do those transactions. Right now it's a combination of me, human consumer, still controlling almost everything but same time relying on collecting information. If you go on vacation. So I want to go to this destination, create itinerary for me, like source some pricing and stuff.

But you're still going to look at that and you put the final stamp of approval and even book and reserve everything yourself because still there is a very little trust into the technology. Right. But eating, right. So like you guys saying so and that one probably still jury is still out. Although in my opinion I think with so much investment into this industry.

Right. So I think eventually we will develop the technology which will help us to be more efficient and improve our lives in terms of the spending less time on shopping and more time on being at the destinations and enjoying services and products.

Steve Phillips [00:20:44]:

Yeah and it's very interesting. So only the other week I needed a new, a new fit sort of, you know, Fitbit type product and I had used Fitbit before and I went straight to—Instead of going to Google I went to ChatGPT and I put in the features I wanted and I ended up with a my new one which is a Whoop and, and it was interesting as a process from, from a consumer who in, in the last time I'd have done it, I'd have said which are the best exercise wristwatches or whatever. And it would have given me a whole series of sponsored lists and there would be communications from those brands directly to me that they are paying for.

However, in the agentic world, or in this case just generative AI world, but it's very similar. I instead put in features and functions I wanted and it told me where and what to buy. Now that's a very different relationship with the consumer. So from a brand perspective, you're talking to these senior leaderships within these brand companies.

How are they thinking about this new world? Are they worried about losing control of the conversation with the consumer? Are they nervous about where the world's going to? Do they think they can, they can survive well and prosper in this agentic shopping world we're seeing?

Michael Nevski [00:21:55]:

Yeah, that's a very interesting question. I think right now I wouldn't say worry. It's more about this is the new approach. It's like in research, right? We very often talk about how we can keep and even improve our seat at the table.

Right. How we can improve corporations because you need technology, new settings. Right. Same idea here that somebody said that it's no longer the search engine optimization, SEO. Right.

It's more about information or informing generative AI. When those engines, like you mentioned, do the search on our behalf and you're optimizing your website to provide, not only now, but provide information which would be exactly what the generative AI is looking for on our behalf. Right? To be on top of those search results, in a sense, or be included in those search results. I think it's more about how we adapt and I know that a large percentage of businesses in general are very much looking at investing in agentic AI.

Right. To enable the network or be able to sell products and services like on the merchant side or issuers, banks, right. When they talk about issuing financial instruments, payments processing. So everybody is giving up right now. It's not about an award, it's like a—I would say half glass full, half glass empty.

Right. It's about an opportunity and how you can prepare better. Right. To enable your business, pretty much putting on a new winter tires, right, when winter comes and be successful.

And that's how I see it in terms of the dialogue.

Steve Phillips [00:23:40]:

That's great. So more away from sort of search engine optimization and more towards information optimization and searchability, which is very interesting. But it does mean that there's the power of the product becomes, it increases whilst the power of marketing decreases. Is that fair? Is that a trend you're seeing? Is that what these people are thinking about?

Michael Nevski [00:24:01]:

I think yeah, at this stage you're right. That's my feeling, again, that's my private feeling being in the industry in general that it's more about optimizing your products, optimizing your content, optimizing your websites. Because yes, you're going to be dealing with the technology, but really pretty much feeding that content to generative AI, whatever engine you're using again to be considered, to be encoded, to be in the information provider. So I think that's the first aspect. Then probably in the future we might see some new innovative approaches on how to do marketing and how to do messaging. So but right now that's the thought process in my opinion. That's what I observe.

Steve Phillips [00:24:43]:

Okay Michael, let's move on to our lightning round where I'm going to ask you a few quick hitting consumer, customer insight related questions. First of all, what's the most unique way you've seen brands use Visa transaction data for creative or innovative product strategy?

Michael Nevski [00:24:57]:

Yeah, I think as an example I gave earlier, the rise of a single person household and single travel. I think that would be a good illustration how cruise liners approach the knowledge of single travelers actually desiring or having a need to travel to international destinations and, for example, tailoring tours and even voiding the second person cabin fees. So specifically targeting single travelers, I think it's one of the good illustrative examples where instead of having a lower occupancy rate and going on a cruise, right, so they can attract single travelers and fill the ship and actually realize not only revenue but also creating that extra loyalty among rising segment of consumer population.

Steve Phillips [00:25:42]:

Brilliant. Thank you. Next up, what's one global spending trend Visa is tracking that marketers and insight teams should know about?

Michael Nevski [00:25:52]:

Oh, very good question. Actually it's still in a nascent stage, but I think and I see that in my work: the generative AI subscription, date subscription, because we all know we have a free versions.

Right? But actually where I see nascent trend of people start paying for generative AI subscription, that means they find it useful. And there are some trends where not only—it's a kind of a double edged sword, I would call it, right—because, let's say, 7% of consumers in Canada already saying that they get a paid subscription. But there is a combination of where they, part of it, they pay themselves and part of it companies, somebody else is paying for them. So I would say we need to pay attention to that trend because I think as it starts unfolding or growing that kind of market penetration, that nascent or early innovators segment will be the first early adopter segment to move forward with agentic AI, with that new innovation, and being prone to pay for products and services which bring efficiency or will do the price optimization and stuff like that.

And I think it's important to pay attention to that trend because it will be indicative or indic—indicative, excuse me—of the adoption of generative AI and agentic AI.

Steve Phillips [00:27:12]:

So that's fascinating. That's fascinating. And it moves us on to the next question, which is what's one area where you think AI will most improve the shopper experience?

Michael Nevski [00:27:23]:

Oh, another good one. Lightning round. Woohoo. I love it.

So I think time spent on shopping probably will decrease, also I think because of the agentic AI. I think second trend, we'll realize that consumers will learn more about the arity of products and services without really spending extra time because agentic AI will be able to come back and educate us, have kind of did you know, by the way, did you know there is a such and such product, such and such service? And start making those very hyper personalized, tailored recommendations for us because we don't know what we don't know. Right. Like from economics classes we know.

And that's where I see the opportunity and the third one probably opportunity as the system or technology helps us to be more efficient. That means we might have extra within our budget which we can allocate somewhere else, whether extra activities, investments or doing something else with that little savings we have. So and I think that's another area where us as brands need to think proactively. How can we accommodate for that extra savings consumers will realize with this technology. So those are three parts I think we're going to see in the foreseeable future.

Steve Phillips [00:28:40]:

Great, that's really interesting. Well, what consumer spending shift in the last year has surprised you most?

Michael Nevski [00:28:46]:

I think shift from physical products and that's what I can see in my internal data to services. So one aspect of that, because your mindset, think of it, American or UK consumer, we're constantly being bombarded by geopolitical events, inflation, some other disruptions, what's going on in the world and people want to feel comfort, they want to take their mind away from what's going on. That means spending more on me time in a sense, right?

Steve Phillips [00:29:22]:

Yes, of course. We need to put the food on the table, we need to eat, like discretionary, non discretionary. Of course, grocery is always non discretionary. It's kind of non negotiable. You need it every day.

Michael Nevski [00:29:32]:

But same time people feel like maybe I'm not going to buy this extra pair of jeans, maybe I'll go and get a spa treatment or maybe I'll travel to Peru. Right. And experience the new culture because I don't want to think about this whole world which is going crushing kind of on me. Right. So that's what surprisingly.

And it's good for some companies as well because consumer shifting more and more spend on service than product.

Steve Phillips [00:29:57]:

That's really interesting. And then finally, do you have a favorite example of a brand using Visa's data to improve creative or innovation strategy?

Michael Nevski [00:30:06]:

That's a good one. It would be a little bit challenging for me to directly provide an answer because I'm not close to creative marketing team. But I think what my company is doing in terms of the supporting the global events. Right. So in working with the companies and brands around the globe, helping them to really focus on what is important for consumers in general, I think definitely there are examples there where by partnering with my company, brands realize some successes in terms of the understanding consumer attitudes and trends and social.

So I think those are very important. I wouldn't just be able to provide you with a very concrete example based on my purview. So. But yeah, that's a great question. Thank you, Steve.

Steve Phillips [00:30:49]:

That's great. Okay, that wraps up this episode of Inside Insights Podcast. Thanks to Michael Nevski, Director of Insight for Visa Globally. Thank you for joining us. If you'd like to contact Michael, you can find a link to his LinkedIn profile on our show notes or at insideinsightspod.com or you can visit his company website at visa.com. If you haven't subscribed yet and want a regular stream of research and insights knowledge in your podcast feed, hit that subscribe button in your podcast app or follow us on YouTube.

Okay, that's all for today. Until next time. Thanks very much for joining.