EP 353 - Why Agentic Commerce Starts With Data: Eli Anselm of Ridge Ventures

0:00 Welcome to the Jason and Scot Show, your source for the latest news and trends in the e-commerce industry. Featuring host Jason Retail Geek Goldberg, Chief Commerce Strategy Officer of Publicis, and Scot Wingo, co-founder and CEO of ReFiBuy. Here are Jason and Scot. Before we start this episode, we have an important announcement for you. Hi, I'm Kiri Masters and I'm excited to introduce Adgentic, a new show about advertising in the age of AI. I've spent the last decade analyzing retail media and now I'm fascinated by what happens when more of the shopping journey moves into AI assistants and agents.

0:48 What happens to advertising when there's no search results page? What will ads in AI actually look like? And where should brands, retailers, and publishers be paying attention right now? Those are the kinds of questions that we'll be exploring on Adgentic through video essays, conversations, and perspectives from the people actually building this new ecosystem. Adgentic is live now as part of the ReFiBuy Media Network. Find the show on YouTube or wherever you get your podcasts and subscribe to the Adgentic newsletter on Substack.

1:29 Welcome to the Jason and Scot Show. I'm your host, Jason Retail Geek Goldberg. And as usual, I'm here with your co-host, Scot Wingo. Hey, Jason, and welcome back. Jason and Scot Show listeners and viewers. Don't forget we have a YouTube channel, everybody. Check that out. Jason and Scot. Today, we continue day two of our live podcasting from Retail Club AI Festival in Huntington Beach, California. We are joined by our guest. His name is Ellie Anselm, and he is an investor at Ridge Ventures. Ellie, welcome to the show. Well, Jason Scot, thank you guys so much for having me on. Excited to be here. Yeah, yeah. You and I first met, I think you called, you inbounded to me and wanted to chat about jump to commerce, like a year.

2:16 It was a long time ago, I looked and it was like in, It was like right when I started Retail Gentic. So I think it was like... Side note, be careful what you ask for. February of 25. Yes. And I was like, someone else wants to talk about this? I think it was like February or March of 25. It was super early in this journey that we're on. And then you guys invested. So that was a... I'm glad I took the call. So that was a good one. Tell us a little bit about your background and then what you do at Ridge Ventures. And we'll jump into it from there. Yeah. So Juan, I'm also very glad and happy we were able to connect back then. A little bit about myself. So I kind of non-linear trajectory into venture, started off on more of the math, data science side of things. So in AI before it got sexy, traditional machine learning in financial services capacity. So it was with the S&P predicting crashing markets, essentially.

3:03 It didn't last very long, given I have too much ADHD to say within a research team and as exciting as algorithms are, had to do something a little bit more oriented around people, jumped shift to the startup side of the world, ironically building software for venture capital and private equity firms. Some thick skin dealing with people like myself on the other side of the table for some time. So both Empathy for Founders did the whole zero to one and building both, you know, product and go to market and that building, you know, company is still off and running. And I shifted to, you know, Ridge roughly three years ago now. I've been on the investment team focusing on, you know, hopefully investing in the best of breed seed and series eight companies. A little bit about Ridge and kind of where we focus our time. So firm's been around for a bit.

3:44 We're currently investing out of our fifth fund and primarily spend all of our time, like I mentioned, early stage. So really everyone from having an ADECAN idea up until Series A companies that might have like, you know, what we like to call early product market fit. All of that is within B2B enterprise software, really like our best of breed across the stack ever from infrastructure to the app layer. And the reason we both focus on enterprise, not from an investment perspective, but also the way we are able to support our companies is what we've coined really is the Ridge Revenue Network. So think of a group of two to three hundred CXOs in mid-market in like four to two thousand enterprises. And one of the main cohorts is actually our CMO council.

4:24 So a group of, you know, CMOs from some of these companies where we, you know, wine and dine them, host events, stay very close, you know, to the nose in terms of what's happening on the ground, where the CMOs want to be investing, what they're looking for in terms of best of breed tools, where they kind of rely on us. And then that becomes a nice flywheel for our portfolio companies as well and kind of feeds our investment strategy. Investors have a lot of different ways they kind of approach markets. Sometimes they'll just take pitches and whatever is interesting to them, you know, is what they'll evaluate. Other times they more proactively go out with a theme Where do you guys fall in that spectrum? I'd say a mix of both, but specifically within commerce, historically have been big investors in like commerce and the MarTech stack.

5:07 So invested historically in, you know, Braze and then on the commerce side, you know, across that companies like Good Day Software, which is AI native ERP. Loop Returns, helping kind of the return system. And even like the CPC brands like Chubbies, which is actually the former founder of Good Day Software. So have experience in this space and me from like a thematic perspective, you mentioned, you know, reached out to you relatively early in this world of agentic commerce. I think before the term really became a thing or I was calling it agentic shopping. Yeah, it was close. I got one word right. 50%. It's OK. You're right. You know, I think you got both words wrong because a lot of it and a lot of what you do is not agentic. So it's true. Oh, boy. Here we go.

5:47 Let's not open that can of worms. But yeah, no, I was lucky enough to kind of be excited about the space, have really good context from just where, you know, the value has been attributed historically. And the way I saw the world going was there is just a, and maybe I'll define like agendic commerce for people in the way we view it. Because I think a lot of skeptics come in and say, Ellie, like this is a great idea, but I don't imagine ever like giving over my credit card, banking information to an agent and like it go buying a house for me. And I'm like, sure, you know, I don't know if we'll ever get there. But the way we view the world is anything within the experience of researching, you know, discovering and then ultimately purchasing products from an AI powered and autonomous experience is really agendic commerce.

6:33 So any part of that stack. So it could literally just be discovering new products is really agendic commerce. And then all the other aspects within that sequence are going to be developed over time. Yeah. I looked at the market and said, OK, for that vision to be realized of people coming in, having seamless experiences and really scaling with like agentic or AI shopping, what needs to be true? And the first thing is consumers are only going to use tools if they have good experiences. And the good experience isn't going to come necessarily from better models. It's going to come from the merchants jumping on this train and saying, what do I need to do for the AI side of the experience to actually be good?

7:14 And that's really where we landed within, I think, product cataloging. But I can talk more about the second kind of how we view it, if that's helpful. Yeah, I think Jason has a question. I love your framing. I 100% agree. And you're going to miss the market view, narrowly defined agentic as that autonomous transaction. But I always turn that question around on people to say, I can never imagine mine now. Because like, could you imagine buying a house without getting agents advice, like to buy? I mean, it's the biggest investment that most people will make. And increasingly, an agent is going to play a role in that transaction. Yeah. So I certainly agree. I'm sort of curious, though. It's funny in my world. One of the biggest problems with this whole agentic thing is the pace of change, like.

7:58 Things are happening so fast that our old models of how we select software or how we craft solutions or how we do change management, they're not designed for this pace. In my mind, and you may tell me I'm wildly wrong, you're in the one industry where the pace of change is potentially an advantage, right? Because it's like you find out if you had a good investment philosophy quickly or not quickly. So, yeah, I don't know. Is it helpful for you that things are happening so fast? Funny enough, in venture, it's probably a hard part of the job. You actually don't really know for some time. So like, you know, maybe early stage, yes, we get a write up on a company pretty quick. But there's plenty of companies, you know, like Groupon that goes off, raises hundreds of millions, and then isn't a great outcome for all the venture capitalists involved. So hopefully, you know, all the companies I invest in have fantastic outcomes, but, you know, we'll continue to see past initial fundraise.

8:48 But yes, I'd say the rate of change actually creates an opportunity for investors like Ridge specifically. There's kind of like bifurcation in the market right now with the mega deals, you know, seed rounds getting done in a billion and the mega funds really taking a large portion of where you could say rounds are in volume, capital is getting attributed to. But now, given that the rate of change is so quick within all these industries, retail and commerce, for example, it actually comes down to, funds like Ridge that are building expertise in this market can stay so close to the ground, understand what's actually happening, and then we can underwrite and make investing decisions faster than like mega funds that are trying to capture the whole world. Sure. Although, side note, I think you did just price a refi by a C round. So just saying.

9:34 What are some of the areas that you're looking at now as you think about where things are going? You've kind of been thinking about this for two years now. What are some areas of interest to you? Yeah, so kind of like I mentioned before, I think of it as a sequence, like to get to this world where everything is natively agentic for consumers and businesses and, you know, shopping, procuring things, what needs to be solved for and be true. So the first layer I really found was to be the data layer where you guys are playing now and enabling all of the agents and AI services to properly capture the end products and SKUs and services that people might want to be purchasing.

10:14 Then the next one and when you talk about all the companies and infrastructure that's getting this intermediated by AI the first layer is actually also advertising and there's a lot of hype going on with ads showing up within AI services so within chat gbt. Obviously they have to generate some revenue to make up for the compute costs and you know their funding rounds and the dynamics you know we don't have to go off on a tangent about their IPO getting delayed for safety but I think everyone is like picking up on on the trend what's really going on but. Actually, AI native ads and offers happening on other surfaces is something that I find super interesting. So travel, for example, has always been limited in mapping inventory, potential offers and promotions that they can give to consumers by one aspect of latency.

11:01 So to be rendered on, let's say, Google Search, for example, and all these other platforms, they have to have a set of, hey, this is what we're able to offer. We're going to push it out. Super low latency systems. You can't actually dynamically, you know, offer anything given those limitations. And people end up going on the site. It's like, oh, this inventory isn't true. The listing for the, you know, room is essentially perishable there. Yeah. And I think because of agents and people's aperture to wait a bit longer for results has now opened a door where, hey, how can we create really dynamic, more personalized offers for all these consumers? So going on to maybe generic site, a side ad, not just being, hey, you know, we know you're looking for a hotel.

11:46 Here's a hotel for two with a hotel view. But they know that your search was oriented around a flight to, you know, this particular country or going with your family. And maybe, you know, they're throwing in a few nights plus dinner or whatever it might be that's oriented for you. So really dynamic experiences on offers and promotions and personalized ads, which I think is very interesting. So you guys have a really nice office right there in San Francisco, kind of in the heart of things. If you look at the wave of kind of companies coming at you, is it higher than ever? Is it like a little bit in the lull? Specifically, I guess, in the categories you're looking for. Yeah. What are some of the trends you see on the inbound of the company formation side of things?

12:30 I'd say in terms of volume of companies, probably within the same pace that it's been historically. If you look at the number of deals that are going on, it hasn't significantly changed at the CD&S A stages. Primarily seed series a and where things have really changed is the bar for rounds to get done now so what was historically maybe when i started in venture 2024 you know the typical series a is a company going from zero to one million and they are in a year and that's like you know right down the middle of the fairway for us very interesting let's go look at the dynamics of the team and the product and market, now companies doing zero to three or you know 5x growth in the year is like baseline.

13:13 And we're, you know, really trending towards companies that are selling into kind of the AI stack, whether it be within inference, whether it be with indirectly to the labs, and the leading token consumers really going zero to 10, zero to 15 in a year is, on one side, like we're really trying to spend a lot of our time and then traditional SaaS, and I'd say like focus AI companies for, call it niche verticals and sectors, you know, commerce, maybe historically So like legal, for example, is more of, hey, what dynamics do they have in terms of underlying data, underlying modes? It's really hard as things evolve to pin what is the compounding value for some of these companies. And we continuously see that.

13:56 Harvey, for example, put out their research work on a in-house model because they're not allowed to use any of their customers' data. You know, questionable about the results and what they're doing there. So I think everyone is grappling around like, hey, even for the companies that are far along, what is really like resounding value? And for the most part, it's actually coming down to just teams execution and speed, like building an expertise in a particular market, getting enough, you know, deeply penetrated products within the largest consumers of whatever it is this tool is. And, you know, getting far and wide as possible within the industry. Speed is the most. Yep. Yep. I know there's a lot of listeners of Retailjantic that are entrepreneurs.

14:37 I'm guessing when you get approached about potential funding, there's some common mistakes that get made over it. Do you have any sort of stock advice for entrepreneurs that are looking to get funded? Hmm trying to think of like the most generalizable yeah yeah for for founders. I'd say on one hand coming into conversations, and first try to get an understanding of the investor's view of the world and of the market, founders sometimes come in with like very deep thoughts and it's great to be opinionated but try to address what the investor is looking to understand like if you see them leaning towards pay explain to me like the competitive landscape don't just say oh you know these competitors they don't know what they're doing you're completely different like, take a second go a quick tour deeper and in explaining like the components of either who you're selling to what you're building and how it's differentiated so active listening and answering questions directly is something that surprisingly like a lot of founders struggle with because they're just so focused on hey this is the pitch i practice and i need to kind of sell, the investor on, as opposed to what can I learn from the questions this investor is asking?

15:45 And like maybe gaps and holes in our narrative. And it's fine to have gaps, like no deal we're going to underwrite is going to be a slam dunk. It just doesn't exist. Like there's always things we underwrite in terms of risk, but just like calling out the risk, being very direct about it. And then your ways of, you know, hindering that risk or hedging it in any way. Metal launched Muse, and then you have Instinct. There's Camp in the Silicon Valley world. It doesn't seem to have as much traction though. There's another one coming that has a human in the loop i forget the name of that one what do you think about this new category of personal agents and are, you starting to see people building on top of that already because you guys silicon valley reacts so fast to these things and in some ways these agents are all kind of a reaction to open claw well they're they're all, they're all derivatives of taking open claw in a different direction type thing yeah i think I think it's all kind of like pulling back to the reality of like the sequence of things and what it's going to take to get to like, you know, native agent of commerce kind of experiences.

16:46 Everyone wants to jump on these ideas. Open claw and like the ability to build these tools and the personal agents with quasi shopping capabilities kind of existed some time ago. But the only thing that's changed is with better models, better connectors, the experience is just better. So people are really jumping on it. I think it has some room to grow. Cute animated characters. Yeah. Those go a long way. Don't underestimate the power of that. They do. Myself included bowing into it. But yeah, I think obviously people within the Bay Area and Silicon Valley jump on them much quicker than others. We do create a bit of an echo chamber in terms of, oh, everyone's going to jump on personal agents tomorrow and the whole world is going to change. But you go to the middle of the Midwest and you mention Muse or even Claude and CIO from a random mid-market logistics company is going to say who?

17:32 So I think it's great, it's a good signal for what we're going to see over the next quarter over quarter growth in both technological advancements and then also go to market distribution. So people are going to start, I think, adopting these tools as they get better, as the experience gets better, and that will ultimately uplift the underlying, market for, hey, what's the infrastructure needed to actually enable these experiences, enable shopping to be completely autonomous. A lot of the things I've seen pulling companies back within Agenda Commerce is actually like the lack of the full stack existing. Like you can build and you can see with a lot of the companies are on payments, we could build all the payment infrastructure for agents in the world.

18:12 But if there isn't enough transactions or overall like agents and, you know, personalized tools for doing the shopping, you don't really have anyone to sell to. So again, coming back to like the sequence of things, what tools need to come into place first, what kind of consumer experiences need to exist and how is that going to scale out in terms of this market coming together and that's why i think like really you need to just give it time someone during the event actually mentioned this is a great i think like analogy for what we're looking at and he's you mentioned like the comparison of mobile shopping to desktop and in reality like it took almost a decade for mobile to really surpass desktop or however you want to you know measure the start and end but, But I think that is like a similar way. But we can look at this.

18:55 It's not going to be overnight. It'll be a slow growing, you know, segment of how people experience their shopping. But it'll be significant. I think it'll take a large share. Yeah. It's funny because I've looked at that exact progression and the like, obviously, everything in agentic is happening faster. Right. So it's fair to apply some acceleration factor. But there's an argument that the jump from non mobile to mobile actually like should have enabled more usage than the jump from non agentic to agentic. Right. Because agentic lets you do things that people are already doing better. It doesn't necessarily create a new audience that didn't exist before. When you gave a bunch of people that like to shop online a mobile phone, they suddenly had way more hours a day where they had the opportunity to shop.

19:43 Right. And so just to put that in context, the iPhone launched in 2006. There was mobile for like 10 years before that, but we'll pretend that didn't exist. The iPhone launches in 2006. Ten years later, 10 percent of e-commerce was mobile. And it was like 13 years to get to 30% mobile, which we're about 50% now. Yeah. And I think you made a good point, which is as you compare the agentic to non-agentic shopping, you're not necessarily getting a new cohort of shoppers, but you are getting a new segment of shopping you can do. 100%. Like there is just certain type of shopping I would not have been able to do previously or would not have done previously if it wasn't for agents.

20:25 For example i went skiing with our group we have like you know with rich has like a annual kind of founder event and i went up i've been snowboarding for some number of years the, you know guy who was taking us looked at the rental board i had he's like man what are you doing the rental board like you're clearly you know, like an experienced snowboarder yeah i'm like i don't know i've never really bought a board like i never thought about it and he's like you absolutely need to buy a board like you cannot go down this mountain doing that yeah and it's always embarrassing to have the rental board on the hella trip. Yeah. Yeah, I was definitely singled out. And I just didn't know what type to buy. I knew that I liked really tight straps. I was somewhere between intermediate and expert. All these nuances and I just don't have the time to go do the shopping.

21:07 But because of opening up a couple of agents, doing some deep search across hours of review videos and explanations for where you are and what you prefer, I was able to buy a board and bindings and boots in like 10 minutes. Otherwise, we're not a purchase. Yeah. And to me, that that goes back to your very first hypothesis that agentic is this broad thing that includes discovery, not just the transaction. Because, you know, we always joke e-commerce saw buying but broke shopping. Right. So like traditional e-commerce got very good at transactions, not so great at like discovering new products and the joy of shopping. And so to me, the big excitement about agentic is that it potentially unlocks. Digital joy of shopping and discovery.

21:49 Yeah. Yeah. Let's we've talked a lot about the front end of shopping. Let's kind of pivot to the back end. Have you seen any interesting AI companies on supply chain or store management or any of those types of areas? Is that even interesting to you? Yeah, for sure. So we've actually looked at it like two, one being actually demand forecasting. So in terms of I'd say like the teams that are really using traditional machine learning algorithms that they would have taking in, you know, data points. That I didn't even know existed in these types of forecasts, like weather and, you know, foot patterns, et cetera, for different geographies and where they're going to, you know, both supply inventory and then plan for on-season and off-season purchasing.

22:29 So on one hand, allowing these teams to do those workflows much quicker, taking very large skill teams into smaller ones and having much more accurate, you know, determinations for where they should be purchasing. And then actually, also, we talked a lot about the consumer experience for ejecting shopping, But procurement for businesses is crazy. And I will tell you, you know, anecdotally, one business is working with, think of one of the top like five largest furniture businesses. They have, you know, non-strategic procurement. So purchasing of things that may be within their warehouses like pallets. There was a company that took just pallets as a category and said, let's go on to market and understand like how you're purchasing across all these different warehouses and, you know, such is really mapped out.

23:14 They identified $4 million worth of overspend, went through to all their suppliers and said, hey, we got these better quotes, meet them or we're going to drop you. And then past that, once they actually secured new contracts, identified that like two different warehouses that are maybe six miles apart were buying from the same supplier at completely different SKU prices. And one of the renegotiated contracts the year following, they tried invoicing them for 500K more than it was agreed upon. And because they just don't have the bandwidth to really, you know, check the details and granularity across all these contracts, all this purchasing are just burning those dollars away. So it's really good to see like intelligence on the spending side, being able to structure all this data in a way that wasn't possible before that I find super fascinating.

23:59 Yeah. Yeah. It's funny. People don't talk about it, but like worldwide consumer spending is about seven B2C is about seven trillion dollars a year. B2B is 14 trillion dollars a year and way more automatable. Like there's it's lower like there's all these EDI transactions. And per your point, like the amount of friction in them is shocking. Yeah, and the lack of tying it all back together because it's been so frictiony. You have the unique visibility of having been at this show both years. So I was there as well. How would you kind of describe it to folks that haven't seen the progression over a year with this show of how it's changed? You've probably attended a bunch of the content. What are some of the more interesting things you've seen on stage?

24:41 Yeah. So maybe to compare year over year, I was really impressed to come back and see the progression. It seems like it's been five years in terms of everyone's like mental model for using AI in this industry where it actually has a measurable return. I think last year it was much more like showboaty surface level of, oh, we've experimented with this and deployed that across whatever it might be, like automated emails and such. And now people are coming in with much more crisp questions and expectations for answers. And how are you actually measuring your AI traffic? Are these numbers that you're throwing up on the board like actual AI traffic or is it 20% consumers and 80% bad actors from Chinese bots?

25:24 So the one skeptical side of things and people knowing how to dissect a lot of the data and things that are thrown up on the board is pretty great because that is ultimately to adopt AI in a better form factor. I also say from the content that I'm seeing, really like how people are breaking down different layers that need to be implemented to open up some of these opportunities. So, for example, Stripe, you know, walking through one, their agentic checkout, you know, saying how most of these agents are failing in the checkout experience. And that's why you're really not seeing a lot of shift into revenues. So they've actually just come out with AI native agentic experience. It's embedded in anyone that has Stripe checkout, which is really impressive.

26:06 As well as like underlying policy and intent. People talking about, hey, like, where is all of the, we'll call it, you know. If the lawyers get involved and people start doing returns and complaining and pointing fingers at agents, who's going to take, you know, the foot of the bill? So not just high level, is AI going to work? Is it not? Like in this actual experience, like how do you solve for some of these different layers, which I think is super interesting. Yeah, we went from theoretical to super tactical in one year. We're crazy. Yeah. Very exciting. Yeah. Well, we really appreciate you taking time. I know you're busy. You got a lot more to go see. You want to take us out, Jason? Yeah, yeah, yeah. So to sum it up, the main reason this year is way better is because I'm here this year.

26:46 Got it. Of course. And I'm invited here. Yeah. And, you know, sadly, that's all the time we have for today. So we're going to have to wrap up. But thanks very much for stopping by and sharing some insights with us. And until next time, happy commercing. You've been listening to The Jason and Scot Show. For all the latest news and trends on e-commerce and shopper marketing, subscribe to us on iTunes or visit JasonAndScott.com.