Retailgentic | Consumer Behavior & Retail Trend

This episode explores the shift from traditional search and ecommerce flows into a new agentic commerce environment where shoppers are using AI surfaces to discover, compare, and increasingly delegate parts of the buying journey.

Ashish explains what Google was seeing internally that led to the development of UCP, why the protocol was designed to be open and ecosystem-wide, and why the ambition goes far beyond checkout. He also shares how Google is thinking about discovery, conversational product data, merchant readiness, and the long-term role of AI across the shopping journey.

For retailers, brands, marketplaces, and commerce tech teams, this conversation is a helpful window into what Google believes is changing now, and what foundations businesses should be investing in before the market fully shifts.

Highlights 
  • How Google’s large-scale ad and data systems shaped his approach to commerce
  • Why UCP emerged now, and the two major trends that inspired it
  • Why Google believes agentic commerce needs an open, ecosystem-wide standard
  • Why the future of UCP extends far beyond checkout
  • How discovery is starting to take shape within the protocol
  • Why richer, more conversational product data matters in AI shopping experiences
  • What retailers need to do now to improve visibility on AI surfaces
  • How Google is thinking about the relationship between agentic commerce and advertising
  • Why agentic commerce is still early, but already moving from concept to reality
Agentic commerce is still early, but this conversation makes one thing clear: the infrastructure decisions being made now will shape who gets discovered, considered, and chosen in the next era of shopping.

Timestamps
2:32 — Welcome and introduction to Ashish Gupta
5:01 — Building Google’s data infrastructure and declarative querying at scale
9:43 — AI at Google before the current wave
11:32 — What it means to be an Engineering Fellow at Google
14:00 — Balancing technical leadership and operational leadership
16:10 — Why Ashish moved into merchant shopping
17:41 — The origin story behind UCP
24:58 — Why Google chose an open protocol approach
27:50 — Responding to skepticism around Google’s commerce efforts
32:30 — Discovery, catalog search, and where UCP expands next
35:39 — Multi-item and multi-merchant cart support
37:43 — New protocol capabilities and partner onboarding
38:42 — Why product data quality matters more in AI shopping
43:22 — How shoppers are using AI surfaces in more complex ways
47:47 — Agentic commerce and advertising
50:10 — Where agentic commerce may go from here
51:46 — Closing thoughts and where to follow Ashish

👉 Connect with Ashish: https://www.linkedin.com/in/ashishgupta98/
👉 Check out UCP announcement: https://www.retailgentic.com/p/flash-googlenrf-announces-universal

🧠 Want to stay ahead in AI commerce? Subscribe and follow along:
📰 Subscribe to the free Substack: retailgentic.ai
📺 Watch episodes on YouTube & Subscribe for updates: https://www.youtube.com/@Retailgentic

Creators and Guests

Host
Scot Wingo
CEO and Co-founder, ReFiBuy.ai
AG
Guest
Ashish Gupta

What is Retailgentic | Consumer Behavior & Retail Trend?

Retailgentic, presented by the ReFiBuy Media Network and hosted by Scot Wingo, is where Agentic Commerce meets retail innovation. We help retailers and brands prepare for the future of agent-driven shopping with actionable news, expert insights, dxis, forward-looking predictions, and practical guidance on Agentic Commerce Optimization (ACO).

Speaker 1:

Talk to a lot of brands and a lot of retailers, and the number one question generally is how can they get discovered on an AI surface? A lot of it is actually to do with this product data and the quality of the product data. When users are coming to these AI surfaces, their queries are getting much, much longer and much richer. And they are talking to these agents sometime as if they are talking to a friend or, like, a sales associate To answer to those queries effectively and match with the right product and the right brand, we need to also have that same kind of richness coming through with the product data which we have. Otherwise, you will not be able to match it.

Speaker 2:

Welcome to the Retailgentic podcast where we explore the intersection of retail, ecommerce, and AI agents, which we call Retailgentic. Our mission is simple. We are shining a bright light on the new Agentic AI shopping megatrend which will disrupt the trillion dollar digital retail market. The stakes are high. Google, OpenAI, Microsoft, Perplexity, Amazon, Apple, Meta, and Anthropic are all fighting to be the Consumer AI Shopping Agent winner.

Speaker 2:

There are gonna be some tough decisions for online brands and retailers to make. What happens to checkout? How should payments work? Should you continue to invest in retail media networks? What does the world look like when over half of your traffic is not from humans but from AI agents?

Speaker 2:

These existential questions are the tip of the iceberg. On this podcast, we're talking to industry experts, analysts, and influencers to get answers to these questions, identify trends, and best practices. This is Retailgentic, your podcast partner in navigating our AI shopping agent future. And now here's your host, Wingo.

Speaker 3:

Welcome to this episode of Retailgentic. Listeners and viewers, we have a special treat for you today. At NRF, this very, humble engineering person came up to me and introduced himself as the head of merchant shopping at Google. His name is Ashish Gupta, and he's the VP and general manager for Merchant Shopping at Google. And he's also an engineering fellow at Google.

Speaker 3:

I had a bunch of UCP questions I asked him, and we had an amazing conversation. It's always good to go right to the source of someone working on this every day, shared some of our feedback for what we really thought would love to have in there and very receptive to to to listen to that. And then I wouldn't be a great podcast host if I didn't throw in there, hey, I have a podcast. He had actually heard of it and is a audience member himself, and I'd love to have you on. It took a while to work through all the timing and everything, but excited to welcome Ashish Gupta.

Speaker 3:

I'm going to do a short intro here, but, you know, we'd walk through his background. We talk a lot about UCP. I get to ask a bunch of questions that have been asked of me about, you know, what do you think they're doing about this? Where do you think they came up with that? So we just go right to the source on this podcast and find out, you know, what they're thinking and how they how they came up with UCP, why, where they see it going.

Speaker 3:

What are some of the foundational beliefs of UCP? And then it's always interesting to hear people's backgrounds because you kind of wonder who who would be building something like this. And it all makes sense, I think, when you hear the story. So I hope you enjoyed this as much as I did. I am not gonna spend any more time between you and Ashish.

Speaker 3:

Enjoy, this conversation with Ashish Gupta, VP and GM for Merchant Shopping at Google. Welcome to this episode of Retailgentic. We have an exciting, treat for everybody. I am joined by Ashish Gupta. He is the vice president and general manager for Merchant Shopping at Google, and he's also an engineering fellow at Google.

Speaker 3:

Ashish, welcome to the show.

Speaker 1:

Thank you, Scot. It is great being here, and thanks for having me. Also, was really a pleasure meeting you previously at NRF, and great that we get an opportunity to connect again.

Speaker 3:

Let's start with your background. So have you been... Always been at Google? Where were you kind of pre Google?

Speaker 1:

So I have actually started with a startup in Boston after I came to The US, and this was a startup in the retail industry, actually, and they were working on logistics and transportation. So that is where I started my whole software engineering career. Eventually, this startup got acquired by Associates, which is now the leader in supply chain management and omnichannel commerce. So I was there for an year or so. And after that, I went to get my Master's in computer science, and then I came to Google.

Speaker 1:

And I've been here at Google now for twenty years and I have won multiple hats.

Speaker 3:

So I'm going go out on a limb and guess you like really complex, large scale distributed computing problems is probably like the area of comp sci that you either focused on in school or you have now found yourself kind of like, you know, really, really fascinated by.

Speaker 1:

Yeah, I have always been very, very passionate about building these large scale, sustainable platforms. And that is what I got to do when I was working with these data platforms. And talking about some of the ad systems, like, actually, I built many of these systems with my team, I took on a different role within Google, which was leading the engineering team for YouTube ads. This was a phase where the businesses was growing, like, very, very, very rapidly. And as part of that, I did lead teams which build the entire ad serving stack, which include, like, the ad retrieval engine, the prediction modeling, all the auction and market space, ads quality, creative generation, asset generation, brand safety, And again, the scale of those problems and the latency requirements and the availability requirements are, of course, very, very stringent.

Speaker 1:

And it needs to work at global scale, obviously. So it's been a lot of fun.

Speaker 3:

At the top, I mentioned you're an engineering fellow. Is that designation at Google? I know we have a big Cisco location here in the Research Triangle Park and they have engineering fellows and it's almost kind of like a tenured thing where you get to just like work on whatever you want. But what what is it? I'm sure it may mean something different at Google.

Speaker 3:

What... What's it mean at Google?

Speaker 1:

Engineering fellow at Google is somebody who has contributed and taken on some of the hardest and the biggest of the engineering problems and help solve them. And they are able to obviously work as an individual contributor, working on different problems. But depending upon where they are, like, they can also, like, be leading teams and engaging and helping build, like, big major systems which are necessary to solve some of our customer needs. For me personally, I obviously currently wear two hats. I do lead the Merchant Shopping organization, but also I am an Engineering Fellow where I am the Ads and Commerce area technical lead.

Speaker 1:

And I've been doing it for several years. And as part of this role, what I do is I help with a big picture technical strategy and technical standards across the ads and commerce organization. So to make sure that anything which we are building has high quality, is sustainable, and how we can continue to push on innovation and just really big picture thinking and long term thinking in terms of technical excellence, technical innovation so that we can deliver the best products for our customers.

Speaker 3:

Yeah. So it's kind of like a classic org chart where you have kind of a horizontal role or a vertical role where you're in the shopping, your operational role, and then have another role in there where you're helping set the standards from a technology standpoint. And so how does it how do you do two jobs at Google? One of those jobs sounds really hard. Like how do you figure out how to do two of those jobs?

Speaker 3:

The matrix work is a little tricky sometimes because sometimes people get in them and they're like, I don't even know who my boss is anymore. So how do you kind of navigate through that?

Speaker 1:

So, it depends. Means, I have gone from like managing large teams to becoming an individual contributor and then again managing large teams. Like, for example, the Ads and Commerce, area technical leadership role, which I am talking about, I actually started that when I was leading the YouTube Ads engineering team. And as part of that, we really thought that it will really be great to have a body of technical leaders who can lead some of these technical standards across the ads organization at that time. So I was doing dual roles.

Speaker 1:

And then as like, obviously, the AI was catching up and there's all this stuff around generative AI, I moved to mostly like an IC role where I spent a lot of time deeply working on this ads and commerce technology. And now over the last few months, I am back to, like, doing two jobs. And, obviously, it's, like, challenge sometimes to work across both and give enough attention to those. But, also, they are also very related. Right?

Speaker 1:

Like, in my in my technical leadership role, I am looking for, like, how we are using AI and generative technologies and everything into everything which we do. And also in my current role as a Merchant Shopping Lead, there's there's a lot of influence of AI, like, especially with Agentic Commerce. And this was something which I was doing even before I took on this role as my hat of, like, the technical lead across ads and commerce. And now it's like I'm doing both.

Speaker 3:

Yeah, I imagine you can almost like in your operating role, you may see a need. You're like, I've heard about this thing we need three times. I'm going to bring it to the group of fellows and kind of say, Hey, is anyone else seeing a need for this thing? And I'm being purposely vague because I don't know what it is, but you know, a giant catalog system that's twice as fast as whatever it is you need. And then that gives you a place to kind of hash it out with people that are seeing something from a different, you know, maybe they're deeper on the ad side or the YouTube side or, you know, some other part of the org.

Speaker 3:

You guys can then look for those commonalities and figure out what the right system is to fit that.

Speaker 1:

Absolutely. We do engage with like fellows across Google. And we do talk about like some of these deep technical challenges and problems and how we can collaborate and work together to solve some of those.

Speaker 3:

Yeah, the other thing people may not, and I think this is still true, but for the folks I talk to, the engineering people I know at Google, what's really cool is a lot of them, you kind of almost get to pick your job, right? It's very much a set up for as an engineering first type of an org where the engineering folks can kind of like figure out what they want to do. And then when you joined on to the merchant, was that something that you were like, wow, this sounds like a really fun opportunity. Did that hearken back to your startup days? You're like, I'm dying to get back into retail.

Speaker 3:

Or talk about what brought you to our fun world out of the YouTube world of looking at what's going on in the shopping area?

Speaker 1:

I mean, first of all, I would say, you are absolutely correct. Like, at Google, there are like so many interesting opportunities all across the board that... And because as an engineering leader, you can contribute in many, many different areas. For me personally, especially as an IC working on ads and commerce overall, I was already digging deeper into, like, Agentic technology and what this whole Agentic commerce world means for some of our systems, both from our infrastructure side, but also some of the experiences side. So as some of these opportunities opened up in, like, the Merchant Shopping organization, it fit very, really, really, really well.

Speaker 1:

And in some sense, yes, I am coming back full circle from my startup experience days back into Commerce. And it has just been very, very exciting to work with businesses across the world now, helping them grow, helping them connect with their customers, particularly at a time right now, which is such a transformative period for the entire retail industry.

Speaker 3:

Yeah. And then obviously, so you joined in August '25 and then January '26, we had UCP. So, you know, give us a little bit of color. Was that something that that you was in flight or is that something that you were playing a role in the creation of? Did, to the extent you can share some of the interesting backstory of, it helps us understand, I think, from an industry perspective, why is it doing some of the things it's doing and to kind of understand a little bit of the backstory?

Speaker 1:

So first of all, just taking a step back, right, like Google has been working on useful AI powered experiences for many, many years now. Like, this includes our product search on Lens. It includes the Try On You feature, which we launched last year as well. It includes, like, the shopping tab, which we have on Search. And some of these experiences are things which shoppers really, really love.

Speaker 1:

I'm talking about UCP in particular. The inspiration for UCP came from two critical trends which we noticed last year. So first critical trend was we are seeing more and more shoppers turning to AI surfaces, which is the AI mode on Google Search as well as Gemini for their shopping needs. And when they are coming to these surfaces for their shopping needs, they are not only discovering products, they're actually going very, very deep. They are comparing different products.

Speaker 1:

They are looking for deals. They are looking for reviews. They are looking for local availability. And as part of that, they are actually making their purchase decisions right on these surfaces. So that is one trend.

Speaker 1:

And as part of this trend, they are also open to delegate some of the parts of shopping to these agents, like checking availability in a local store by just having the agent make a call. Second trend which we have been seeing is the Agentic technology is finally catching up, where agents are able to take some of these tasks and do them pretty effectively on the user's behalf. So now what we see is the users are ready to delegate some of these tasks to the agents. The Agentic technology is there where they can complete some of these tasks, but the underlying layer which allows these agents to talk to the businesses was not there. And this is what was, like, some of the inspiration for us to think about, like, how do we build that connection between the agents and the businesses?

Speaker 1:

And as part of that thought process, like, we knew we wanted to do something. And there were a few things which we wanted to make sure the solution entails. Right? Like, the first thing is it needs to work for everyone. It needs to work for all the small businesses, all the large businesses.

Speaker 1:

It needs to work for all the payment providers. It needs to be basically be agnostic to any single platform, any single company, and it needs to work for the entire ecosystem. So that was one thing. The second part, the thing is, like, while we talk a lot about checkout, right, as part of these agentic actions, for some of these actions to work and for the agent to be truly successful and useful to the shopper, it needs to cover the entire journey, right from, like, discovery to purchase, to even post purchase action. And it should be extensible to even other verticals in commerce, not just shopping.

Speaker 1:

So these were some of the thoughts and aspirations which we have when we started thinking about, like, we need to do something here. At the same time, we also noticed, and this was sometime in early October, by the way, when we were having like these senior leadership discussions within the Commerce team talking about like our strategy around AI services and how to deliver more value to the shoppers and to the business. And we saw this gap overall in the industry. At the same time, we did notice that there are, like, other solutions which were coming up in the industry as well. But what we saw was that some of those solutions were, like, very much specific to a particular platform and did not solve the need for the entire ecosystem or for the entire commerce journey.

Speaker 1:

So from there, we started exploring this further. We had started having like some discussions with our industry partners, and as part of that, when we discussed some of these challenges and some of the initial concepts and initial ideas, it was very clear that we were on to something. And from there, we started defining some key principles of what we want to do here. We wanted to build something which works for the entire ecosystem. It needs to be grounded in user value.

Speaker 1:

It needs to be secure and private. It needs to be grounded in merchant value. Even with transactions, we want the merchants to be the seller of record, for example. It needs to support like the existing standards which are in the industry, like A2A, MCP, AP2. And again, it needs to just work for everyone.

Speaker 1:

So we started with these standards. We took some of these proposals then to the Google leadership. There was a lot of excitement that there is a need here to build something, and from there we engaged very closely with our partners build this protocol. And we were so excited to just announce it and share it with the rest of the industry at Nerf.

Speaker 3:

Yeah, how did you get Sundar on board? It seems like, yeah, it was, I knew something big was happening when I saw that he was coming to Nerf. I was like, this is, I've been in this industry forever. I was like, a a senior Google leader of that stature has never never been here. So something something big is going on.

Speaker 1:

So we were obviously very, very fortunate to have Sundar be at such a large stage at NRF, announcing UCP and sharing it with the rest of the world. So again, like, since we started, we obviously had this in the mind. We are building something for the future. We obviously wanted to make sure that we build something which is long term and sustainable with very, very strong principles. There's a lot of work which went into, like, building it with the best technical standards and ensuring that there is, like, strong sense of excellence in everything which which we do as part of part of the protocol.

Speaker 1:

And we were just fortunate that everything lined up pretty well with Sundar on the stage, we were able to announce it.

Speaker 3:

That would have stressed me out. You you seem you seem very unflappable, so I'm glad I'm glad you you were running that. I would have been stressed out. I

Speaker 1:

would just say our teams here were and our teams here at Google and also our partners, especially the foundational partners who we worked with, they were, like, extremely inspired by the vision which we were taking on. It was very clear to all of us. Right? Like, Agentic Commerce is, like, we are building something for the future of commerce overall. So the aspirations and the ambitions was truly big, and obviously, having the opportunity to announce it at a stage, so Biggest NRF by Sundar was obviously one of the motivators as well, and the teams worked incredibly well and incredibly hard, like, with the partner teams.

Speaker 1:

There was incredibly great collaboration across the board with people working, like, continuously on, like, exchanging ideas, exchanging concepts, exchanging, like, code reviews and everything and fixing issues as we saw, debugging issues, and just going from there.

Speaker 3:

Having been in e commerce for so long, it's exciting to see standards like this come up where you publish them into the open with a foundation and basically, give them to the community. The old school way is, you know, pick whatever proprietary company would come up with an API for their marketplace. And it was super proprietary and you couldn't tell anyone about it. And then it was metered. You know, the products want to be listed and sold and they're, they're, they're creating these friction points that don't necessarily need to exist.

Speaker 3:

So it's exciting to see, you know, the way in the current world where we're building these things more out in the open with a way more spirit towards contribution and a shared amount of energy put towards it, which I think will end up with it becoming a standard, much more, you know, higher probability of it working out when you do it that way.

Speaker 1:

Yeah. And this is actually really, really critical for the future of Agentic Commerce also, that we build something which is open and it can work for everyone. Like, if you build it in a closed way, then it cannot solve the needs for the entire ecosystem. And that was one of the principles which we started very, very early on, that it is not something which we are building just for a single company or a single agent. It needs to truly work for everybody.

Speaker 1:

Because, again, from our perspective, right, like, from our... A role of of Google, we consider ourselves as kind of a matchmaker. We want to connect shoppers with merchants and businesses to satisfy their needs. And as part of that, especially with all the new AI experiences which are coming up, there are so many new opportunities where the agents can, if with right infrastructure and the right foundations, can deliver so much value to the shoppers and the businesses. So investing in something which is so foundational with that goal and the mission was something which we really, really cared about.

Speaker 3:

Ever since January 11, the naysayers have come out and said, Google's tried this, you know, pick a number. I think it's actually four. I went back and trace this, but they'll they'll usually use a bigger number. Like they've tried it 10 times and they've always given up. And I was actually intimately involved in these.

Speaker 3:

This predates your, your probably time even at Google. But you know, there was, there was one which I think was called Google Shopping Marketplace. And then there was, you know, buy, buy on Google, think is what the other one was called. And there were like these attempts at kind of replicating more of an Amazon style marketplace, you know, kind of the Google Shopping catalog is here and we're going to hang some things you can buy off there. And Google is the merchant of record type of, the old school marketplace way.

Speaker 3:

How do you, you know, if someone were to ask you that question, how do you kind of say, I have my own answer and I don't believe that, but it's always, it helps me to defend, to kind of help people understand. I understand why the naysayers say that, but I think the unintended consequence of that is it gets a lot of retailers kind of taking a wait and see. And I don't want retailers to get caught behind the eight ball by doing that erroneously. So how do you kind of answer that criticism?

Speaker 1:

So if you look at, like, how the shoppers are coming to Google today, right, like, people across the world are coming to Google with their shopping queries over a billion times a day. And there are millions of millions of businesses who are trying to serve the needs of these shopper with the products and services which they have. And as I said, for us, the key thing here is to be a matchmaker between the shopper and the business. And what's happening now with AI and the AI revolution is that traditional ways of shopping is changing dramatically. And there are new AI experiences which are coming up, and the shoppers are embracing these new AI experiences.

Speaker 1:

And what is critical in these new AI experiences is for these AI agents to be able to connect with with the business. And that is where the UCP launch and some of the foundation building blocks which we are building come in so that these agents can talk seamlessly to these businesses and we can deliver, like, really more delightful experiences to the shoppers. And also from the business perspective, it allows them to connect with these shoppers where they are as part of these new experiences. And from a checkout perspective, even though we started with checkout as the initial capability for UCP, the ambition is much, much bigger and much broader. We want to bring some of these agentic capabilities throughout the shopping journey, starting from discovery and obviously purchase, but even post purchase, like, where is the order or what other actions you can take on the order?

Speaker 1:

So the ambition is much, much bigger.

Speaker 3:

My answer is usually Sundar came to Nerf and announced it. I think they've got senior leadership support on this one. That, yeah, that never happened with the previous attempts there. I do think it's interesting. So you said a billion times a day.

Speaker 3:

I know that's not a billion users, so it's a billion queries, a billion kind of product related queries a day. Okay. Just want to make sure I understand them. Yeah. This is what's nice being at Google and sitting on these platforms.

Speaker 3:

I forget you guys have like 10 to 15 platforms that have multiple billions of users using them all the time. Do you know that number off the top of your head? It's a big number. And between maps and Gmail, just like all the different things that you guys have. And there's almost like a, it's interesting to be able to sit there and see the pull in the market.

Speaker 3:

So you're not just like randomly dreaming stuff up. You're responding to, in this example, with Agentic Shopping, you're responding to what the users are actually kind of demonstrating through the intent they're giving to you, which is the form of a prompt or a search query.

Speaker 1:

We obviously have lots of products with billions of users coming to those products every day, and as I said earlier, even when we think about like some of the inspiration from UCP, it came from looking at some of the user trends which we are seeing on our platforms as more and more users are coming to our platform and going to the AI surfaces for their shopping needs.

Speaker 3:

I could spend another two hours talking about UCP. You know, an area, I've talked so much about the checkout. I think it's kind of like, you know, and you probably have to, but I think we'll of skim over it because it's like, it's probably the most known, most talked about part. And I think some of the other parts are interesting. I'm kind of interesting about the discovery part of it.

Speaker 3:

And, you know, so... And I'm going to splice this in, but you guys have kind of a block diagram. I'm sure you haven't memorized, so I don't even have to pull it up. So in there, there's like, there's blocks that are solid, which are the ones that were kind of you're announcing, and then there's some coming soon. And I don't really wanna, get you in trouble by pre announcing anything, but if you happen to, that'd be great.

Speaker 3:

Just kidding. I'm interested in the discovery because I spend a lot of time on the product catalog part of the world. And it seems like at some point when everyone's lit up a UCP endpoint for their inventory, it seems like you're going to need like a to maybe walk up at least have like, if I was building a shopping agent, a vertical shopping agent for shoes like shoe collectors, I would, I would want to like ask someone like, Hey, what's a set of UCP endpoints that where I can find shoes or Nike shoes or, you know, something like that. Is that, is that part of the, like, where do you end? And this is another interesting question, like building a system is like, at some point you have to end right or else you can't do everything.

Speaker 3:

What's kind of like a philosophy of thinking through kind of like, do you stub out and say, someone's going to to go build that that's not us?

Speaker 1:

So in terms of, like, UCP, obviously, we started with a checkout, as you said, but the aspiration is always being to cover the entire shopping journey, right? And the second thing is we obviously open source the protocol so that we can get contributions from the entire industry. Because, again, it's not only about, like, whether we can just build everything by ourselves, but it's actually important to have the industry contribute to this so that we get all the nuances right, right, because there's so much nuance into... In retail. The journeys are so complex, and there are so many differentiated also that we need all the flexibility and customization which necessary in the protocol so that we can have these agentic actions work at scale.

Speaker 1:

Now talking about discovery, we obviously want to add other discovery capabilities to the protocol and we actually already started adding some of those now. We already have in the protocol already on GitHub Live where there are capabilities to do catalog search for a given query. You can look up a particular catalog, and we are working very, very actively with our industry members on how can we evolve those initial capabilities which we added for discovery even further in more different ways.

Speaker 3:

I saw that get added, I think, in the last two weeks. One question I get a lot, sorry to pick on back over to checkout now, but is, you know, right now checkout is single item, single cart type of a thing. And I'm sure you get this from merchants like, you know, we have merchants in the beauty category that an individual item is $5 which doesn't really work for a cart, but their users on their website throw 10, 15 items in the cart and to get a basket size up towards like $75 $100 that part of the vision of what you guys see executing on it at some point in the standard?

Speaker 1:

100%. Actually, when we talk about, like, a single merchant multi item checkout, the protocol already supports it. As part of the checkout, you can already add multiple items. Now there's a different question about, like, agents being able to support it in terms of the experiences as well as the merchants having the capabilities to handle that. But the protocol itself already has support for this single merchant multi item card.

Speaker 1:

And when we talk about, like, a multi merchant also, like, that also comes up a lot. Like, what about multi multi merchant, multi item card? Again, the single merchant, multi item capability just extends into it. When you talk about multi merchants, there is no two phase commit kind of a protocol, obviously. Right?

Speaker 1:

There is no... Nothing like like the checkout happens on all merchants or none. They are all parallel things and parallel actions, and that is supported by the protocol.

Speaker 3:

Yeah. When I asked the question, I hadn't recorded the the the podcast that came out on the twelfth, March 12 podcast, which was these guys I found in Australia called Villa Labs. And, they had it... They had they had put together a multi item, multi merchant cart with UCP, and they demonstrated it live, which was pretty cool to see. You'll have to I'll cut that out and send it to you.

Speaker 3:

You'll enjoy it. Vendee, Vendee Labs, apologies, V E N D E. Yeah. So it was pretty cool to see. They did shoes in earnestly enough.

Speaker 3:

Everyone loves a good shoe example. What are some of the other things, you know, we're actually releasing this the day you have some news coming out. Walk us through some of the things today that you guys have announced.

Speaker 1:

I think the one of the major things there is obviously the extensions to the protocol itself, where we are adding new capabilities into the protocol like discovery so that it can start addressing some of the more upper funnel journeys in shopping. The other thing which we are also looking into right now is how can we onboard more and more of our businesses onto the Google platform and how can we integrate them with our AI services. As part of that, we are starting to work very, very closely with many of our partners across the industry. This includes, like, some of the very big names in the industry, and we are partnering with them very, very closely so that we can onboard more and more of the businesses onto the Google implementation of UCB.

Speaker 3:

You and I have had interesting, you know, I'm a data feed nerd because I, you know, once you live in the Product Catalog land, then you spend a lot of time in data feeds. And think I it's really interesting having been in the data feed industry, there's kind of like legacy data feeds, which are pretty, you know, they're purposely kind of narrow from an attribute perspective because there was this weird, weird time before AI where kind of the more attributes you gave, it would sometimes tip the shopper journey kind of over into a weird situation. So so there was incentives aligned such that that, would only send kind of the minimum viable attributes, if you will. And it usually is like under 10. So somewhere between five and ten attributes.

Speaker 3:

Now, in this new agentic world, that is totally flipped to not only do you know, LLMs love, they want more content, but they also want more context. And content is, let's provide every possible attribute you can think of so they have it, you know. When I'm talking to a retailer, I'm saying, take your PDP and you know, if you printed that out, maybe it would be three pages of information. And a lot of that's window dressing. You could probably boil it down to one page of text.

Speaker 3:

And the engines want like 80 pages of information. Like, you know, the more, the more you can give a LLM to kind of chew on and figure out all the connections to then tie back to the shopper's context, it will result in a better shopping journey. So you guys have done some cool things there. Talk about Google's view of this and what you think, if you were talking to a retailer, how would you explain why do you guys need this information now? And then some of the ways you guys are going to provide to retailers or are providing to provide this information?

Speaker 1:

So, Scot, this is a question actually I get very, very often. I talk to a lot of brands and a lot of retailers, and the number one question generally is how can they get discovered on an AI surface? And for us, a lot of it is actually to do with this product data and the quality of the product data. So as I mentioned before, like, what what we are seeing is when users are coming to these AI surfaces, their queries are getting much, much longer and much richer. They are going into a lot more details about the product they are asking.

Speaker 1:

And they are talking to these agents sometime as if they are talking to a friend or, like, a sales associate at a physical store. So the queries are getting very, very nuanced. And to answer to those queries effectively and match with the right product and the right brand, we need to also have that same kind of richness coming through with the product data which we have. Otherwise, you will not be able to match it. That is where we are investing a lot as well.

Speaker 1:

We have recently, at the same time at NRF, we also announced conversational attributes, we added to our product feeds. We already actually get a lot of product data and product listing data from our merchants. In our Shopping Graph, we already have, like, over 50,000,000,000 product listings. They get updated continuously with, like, 2,000,000,000 of these listings every hour, and making that data even richer helps us respond to the user queries. So we added like attributes like typical questions which people are asking about the product and what is the corresponding answer to them, or what is the theme around the product, or what are some of the related items which go along with this product or substitute items which go for this product.

Speaker 1:

And having all of this information helps the AI model to find the right product and the right brand given the user query.

Speaker 3:

Yeah. And I forget when you guys put it out. It was sometime fall of last year when you introduced it, and it's just gotten better and better over time is the, it's the it's the view in AI mode when you're looking for products. And for some reason, it gets a signal that you may want to compare them and it puts the little checkbox next to it. And then you you hit that and then it will it will...

Speaker 3:

Now you're in a conversation about those X number of, I'm sure there's a limit. I've gone like four or five. Anything beyond that's like maybe too much, but I really like that mode because if you have, I had an experience buying a coffee machine, kind of like a step above a cup type, but like not, not a Ferrari Italian thing that's super complicated. And I couldn't figure out amongst the four or five I was considering which one had like, you know, cold brew, you know, and a couple other features I wanted. And that mode is like super helpful for that where you're kind of like, and then you can, as a shopper, you start to learn, well, what do people, you know, what do they dislike about this?

Speaker 3:

Just so you can kind of like at least know, is it, are people complaining that it's too hard to use or is it like some super coffee person that needs that Ferrari thing and it's not a fit for them, which I'm okay with. So it seems like that wouldn't work. That level of comparing these things if you had just this sparse information would kind of fall over if you didn't have this type of information.

Speaker 1:

100%. I was actually in the market also to buy an automatic coffee machine, by the way. My requirements were pretty simple. I wanted something which can make good lattes and easy to clean. And I went to the AI surface, and it gave me a couple of recommendations.

Speaker 1:

I compared and decided on the spot. But to your point, as some of these queries get a lot more complex and nuanced, you need that level of granularity in the data which comes with the product, so you can actually match the right product with the right user and the right query.

Speaker 3:

And now that you've had a lot of these surfaces out, you kind of started the story of you were seeing this like inherent demand, and now you've started to try to help with it. Do you start to see now users are like, they're pushing it and they're kind of like, wow, that they're engaging with it. And now they're kind of like pushing it even further. And do you, is that kind of how you guys find product market fit and keep tuning things to this? Like, have you seen anything interesting?

Speaker 3:

You know, I don't need numbers or anything, but directionally that you've seen there as you've... It kind of... It also could be a self fulfilling prophecy too that, you know, now you've given them a place to go, maybe even more have kind of like said, wow, this is pretty neat.

Speaker 1:

Yeah. As we are adding like new capabilities, like especially the new AI capabilities, we start people using it like new and different ways. Like when we launched like Try On You, now you can see like a lot of people are engaging with it in very, very interesting ways already. Right? And similarly, as we are bringing in these new agentic capabilities, it is changing user behavior.

Speaker 1:

And when we talk about agentic commerce in particular, right, like, we all know, like, we are still in a very, very early days in terms of agentic commerce. There's a lot of work which needs to be done for Agentic Commerce to work at scale across the industry. But all the work which we are doing in terms of laying these foundations allows for many of these new experiences to show up, and as they are showing up, people are embracing them. People are liking them. So for me, it kind of a flywheel.

Speaker 1:

Like, as you are adding new capabilities, people are embracing them, which also encourages us to add new capabilities as part of this, and which, for me, like, especially thinking about long term, while Agentic Commerce is still very, very early, it's more of a question of when, not an if, in terms of, like, how this trajectory will go. And, yeah, learning from some of the experiences which users are using and then iterating on this is basically what we do.

Speaker 3:

Yeah, we did talk about how you're the fellow of ads and commerce. So I'd be remiss. I'm sure in the back of your head, you're always thinking, you know, you've worked a lot on the ad systems and things. And I know Google has a couple of ad formats you're experimenting with, and there's a lot of experimenting just in the industry around how do these formats work? Someone's going to ring the register on all the costs associated with this down the day.

Speaker 3:

And a lot of merchants may want this. When I talk to brands, their number one question is, this is awesome. How do I invest in this even more and make sure people are seeing that I support it type of thing? So I'm sure you get that a lot too. Where do you see kind of this intersection of Agentic Commerce and ads happening and how do you, you know, I know the answer to this, but Google tends to be very purposeful in these, the way you build things.

Speaker 3:

So I'd love to hear your kind of take on that to how you guys approach this.

Speaker 1:

There are definitely a lot of possibilities here at the intersection of ads and agentic. But just to be clear, right, like, my team currently is very, very much focused on how to bring some of these Agentic capabilities onto the organic surfaces, making sure we get them right. And now talking about Ads, right, more broadly, Ads has been using AI for a very, very long time to deliver more optimizations for the business, deliver more efficiencies, and as these new AI capabilities are coming up and the power of Gemini model and everything, we are delivering even more of those capabilities. And when we talk more broadly also about what the advertisers and the businesses are looking for, what they are looking for is how can they connect with their customers and how can they grow their business, how can they have new relationships with new customers and everything. And as part of that, obviously, with the new AI experiences, we are always experimenting with new formats, new ways in which businesses can connect with the users.

Speaker 1:

But the main thing here is we need to make sure the businesses are getting the foundations right. As part of those foundations, making sure you get the right product feeds, which we talked about, making sure you are allowing for more automation as part of your campaigns so that the agents and the AI can connect and match the right product with the right shopper's needs as well as like building some of these agentic capabilities for the future, right. So those are some of the foundations which advertisers can continue to invest in. And as the entire ecosystem evolves, there are obviously a lot of new ways to connect these shoppers and these businesses.

Speaker 3:

You seem to be able to handle lots of things simultaneously. So you live in a tactical world where you're implementing this. So look a little bit out in the future. Where do you see Agentic Commerce going? Is this agents talking to agents on your behalf?

Speaker 3:

Are they predicting things? Where do you see this going?

Speaker 1:

It's very early, Scot, from my perspective, to predict where all Agentic Commerce will go because things are changing very, very rapidly. What I can say, though, is it has gone from just being a concept to more of a reality. We see more and more users are embracing conversational AI. They are ready to delegate some of the tedious parts of shopping to these agents. We also see businesses embracing this new world.

Speaker 1:

So there are lots and lots of opportunities for us to continue evolving the entire landscape. And what, again, is exciting is, like, the entire ecosystem is evolving very, very fast in this area. But I don't want to predict, like, exactly where it will go in, like, three years from now or five years from now, because given just the pace of everything, how things are moving, things are changing very, very drastically.

Speaker 3:

You've been very generous with your time, which I appreciate, but I wanna make sure I I get you out before the top of the hour. Ashish, thanks for coming on. It was a delight to meet you at NRF, and I hope to see you at another trade show at some point. Sadly, I don't think you're gonna it to Shop Talk, but maybe I'll I'll see you at something in the fall or or next year at Nerf. We'll we'll, we'll get to hang out a little bit.

Speaker 3:

Appreciate you coming on. Any last words? Do you do you publish things frequently, or are you more a behind the scenes type person? You're not you're not on LinkedIn or are the Twitter sphere throwing your thoughts out there? Where where could...

Speaker 3:

If... Do you publish kind of stuff at all, or are you behind the scenes? And if you do, where should people go look?

Speaker 1:

Scot, first of all, thanks for again for having me. It was a pleasure connecting with you again. I do sometimes post on LinkedIn and also on Twitter, but but, yeah, I'm very reachable for... On LinkedIns. And especially in this current world of Agentic Commerce, we have to build this future together as a community, and I...

Speaker 1:

We are always looking for feedback. I'm always looking for feedback on what we can do more and what we can do differently. So always looking forward to connect with people and hearing where we can do more together.

Speaker 3:

Thanks for coming on, and we'll talk to you when you have some more news to share.

Speaker 2:

Follow, rate, and share to stay up to date with Retailgentic, your guide to the future of AI shopping agents.