Latest / Talk Commerce / AI Readiness and Agentic Commerce for Brands with Amera Khalil of Feedonomics

Transcript

AI Readiness and Agentic Commerce for Brands with Amera Khalil of Feedonomics · Talk Commerce

• [00:00.0] Hi, everybody. My name is Amera Khalil. I lead the Strategic Account Management Organization at Commerce, which is brand that is both Feedonomics BigCommerce and MakeSwift. Yeah. Tell us a little bit more about each of those brands and what they do. Yeah, it's amazing. So I've been at the company for six years and primarily was supporting Feedonomics, which is one of my favorite topics to discuss, especially around the AI era as it relates to feed management. [00:22.6] So we like to say in a simple way, we're taking data that's a little bit complex, we're making it easy and beautiful, and then we're piping it out to all of the different places. We like to say it's not dumb pipes. They're quite intelligent, and they're allowing us to be very strategic and nimble for many brands in the space who have a lot of endpoints that they need to decomplicate or areas that they need to have a little bit of less friction and then BigCommerce is a fantastic place because it is a E commerce platform. [00:46.5] And so we've seen a huge number of people coming to us because we can help support maybe some of the not cookie, cutter experiences that they're trying to create. And BigCommerce is able to support with that. And then makeswift is something that I think is one of the most exciting things as we talk about the AI discussion because we are able to leverage a very nimble and agile, system that makes something that maybe is traditionally not so pretty in the back end beautiful for the consumer. [01:09.2] Consumers when they're searching on various websites. Yeah. Well, great lead in to my next question which is going to be how is the E commerce industry changing nowadays with the advent of AI? Well, being here at Shop Talk is fantastic because that is exactly what we've been talking about. I have people coming by every second and they're like, whoa, AI, AI everything. [01:27.8] And in fact we're right next to the AI stage. So I would say that E commerce is changing and pivoting in terms of consumer behavior. We are seeing this massive shift where people are taking kind of the expectations on like Google performance spending ads against your Google performance. All of the things that marketers once did, tracking, performance lift, all of the kind of traditional funnel that we once kind of made and controlled and that's being collapsed, in an appropriate way that allows for people to leverage AI related search queries, AI related of consumers. [01:59.0] So let me ask you a question. Do you use like Perplexity or OpenAI or Claude? Yeah, definitely. Yeah, right. Everybody. We haven't heard a single person who's like no, I don't have that downloaded. Yeah. And you know, you probably use various LLMs for like behaviors, right? Like I use OpenAI to figure out like my emotional needs, like hey, I've got this tough conversation, all these things. [02:19.7] But I go to perplexity to then, hey, I'm, you know, buying this suitcase. Which one's the best for mine? So like we're not unique. Everyone is doing this. And when we think about that in the era of what we're talking about from the biggest change in E commerce, it's the preparation to be visible on these LLMs while maintaining the quality of your data on any of the E commerce platforms that you know. [02:38.5] But then furthermore being able to see the landscape of strategy that you need to be able to like be nimble and move things quickly to Help really move where your consumers are. And then the biggest thing is that our newest consumer is a brand agent who's kind of searching on behalf of your search query. [02:54.6] And that's a pretty big challenge for a lot of people to manage, both from the E Comm platform, but then from the actual paid ads and kind of these incredible marketers who had this system that is no longer really working the way it once did. Definitely. And so my next question, I'm sure you could write a book on this, but maybe you could distill a few pieces of wisdom for us. Sure. [03:13.0] How can brands and merchants prepare for Agentic specifically? I love that question. You know, I think it's evolving if I'm being honest. I think one, it's not a one size fits all approach. So what we're talking about in terms of just like standard foundation, even about pre agentic standard foundation is leveraging, you know, quality data. [03:32.2] Right. Your.com might not have everything in the right format when you're sending it to any of the channels that you're exporting your feeds to, which is a really standard practice. So having a feed provider, somebody like a Feedonomics who can, can come in and bring in the necessary details, brand title, size and description, things that kind of aren't the most attractive topic. [03:50.3] That's like the foundation that's totally imperative to be successful because you need, when the search query is happening, you need to have your information readily available, for the LLM to pull it. But I think the bigger part is figuring out what your business strategies are. You want to be able to have an appropriate partner that can help with trackability so that you're deciding on how you're going to a B test and send certain feeds to one LLM because we know that they're not going to be the same. [04:13.2] Thinking about the audience for Copilot is very different than perplexity. Right? So like that part is definitely like you need data tracking, you need performance tracking. But then the bigger part is enriched data and enriched here at least I've heard it throughout the week is a little bit of like a buzzword. [04:28.8] Everyone's like, oh, we're enriching our data. And I think one of the things that I'm learning is that enrichment can be different for each, each brand, each category, each channel that it's going to. But it's taking in some of the heavy behaviors that requires a person to manually do a lot of work. [04:43.9] Example, changing the copy on a product for holidays, somebody's like actually doing that manually. Well, we can leverage enrichment, Data Enrichment, one of Feedonomics products that can then go ahead and complete that for you in real time, like very quickly, so that you're just a lot more agile. [04:59.9] That's like a baseline. But then it can get even further, making sure that the right expectations of these long search queries, again, not the caveman search. Right. I'm buying a suitcase, more of like, I'm going on this trip and I want it to be light. And I tend to pack all these things. Right. Like, that's the consumer behavior now. [05:15.8] So it's pulling in all of the necessary data by making an enriched feed file so that when the LLM receives that search query, then they are making sure that the right information is being served. Because if they send the bad information, the consumer's not gonna come back to their LLM. Right. And enrichment is imperative to be able to do that. So we're thinking about AEO and geo. It's important. [05:33.8] Very cool. And yeah, aside from this podcast, my day job is in content and writing and video. And so we've been on the front lines of making those kinds of changes for, say, holiday copy. And we've seen that AI can save you a lot of time, but you do have to be careful because there's the whole humans in the loop element. [05:52.6] You need to control the quality. And, I think there's also some risks that come along with AI as well. Maybe you could speak more broadly to what risks that brands should be aware of as AI becomes more and more part of the picture in the workflow. Well, we haven't got it all figured out yet. [06:09.2] I've heard on stage, I was at Etail west and there were some incredible people and they said, we're not doing AI for AI sake. And I love that. I thought, oh, yeah, you're right. Don't just put AI out there, because what is it going to do? And so I think there's this notion that we still need to be able to provide quality assurance. [06:25.1] We still need to be able to have guardrails, that is assessing the AI, whatever it might be. So if it's in the example, we're talking about completing a new romance copy, being able to go in and write it. We have systems, and Phenonomics is really incredible with this, where our, data enrichment system allows us to benchmark, our own ratings. [06:42.7] Meaning, like, hey, we're okay with this much information, or we're not okay with this much information, and I can give you like a simple example, meaning that, like, you know, we want it to have, a check, a credibility check, and then it's going to give you a scale. And so the nice part is that I think we have a long way to go before AI is perfect. [06:59.9] And I don't know if it'll ever be perfect. Right. But having the human element, because there is a world where there's fear around AI, Right? Like, it's going to replace all these things. And I don't necessarily see that. I see it as an additive and a supplement to reduce friction, to allow speed, agility. [07:15.2] Have it be so that brands can take ideas and speak about it within their internal leadership and then push it along faster and execute. As long as there is the ability to have the quality assurance and the human element that is monitoring the behaviors. Because, you know, when you can read something that's AI, right? [07:31.4] Like when you get an email from somebody, you're like, oh, that's AI, you know, so we want to be able to remove that because we know that there's different, lenses. I actually saw the Gen zers are very different than millennials in terms of what their expectations of AI are. They have feedback. They're like, hey, we don't want it to be, so scripted, and they can tell when it's kind of cold. [07:49.8] AI behaviors. So being able to adjust the content based off of the changes of your audience is going to be really important. But all of it comes back to quality assurance. So maybe we could talk a little bit about some customer journeys that brands have had as they adapt to AI. Yeah, that's a really great question. [08:05.9] I think, you know, at Commerce, we're really lucky. We've got some of the most, incredible brands in the industry and they come to us with a problem. And their problem is, I've been told by the business or the board, I need to have every single feed that is related to my brand. So examples could be Dell or Tapestry brands where they come to us and they say we need to be everywhere and we have to help kind of dissect what that means in terms of implementation. [08:27.9] And so one of the things that I think we were really fortunate to do is, of course in this kickoff, we were running beta tests with these clients where we were not only completing some of their actionable requests, meaning how do we get live on OpenAI? How do we get live on Perplexity? How do we make sure that our data is enriched to the best practices on a channel by channel basis, but then our mapping to the schema of an LLM that doesn't necessarily have clarity on what is good. [08:53.0] Right. What does good look like? That's one of the questions we bring up is like, what does good look like? Well, we're learning. And so we have these asks and so they come to us and we're able to then kind of put together what we call our roadmap strategy. And it can be as complex or as simple as, okay, Dell, we're going to put, a request out there to get a relationship approved so that you are going to receive a feed. [09:12.3] So for some, like, breakdown that I think is really important is just because you're a brand doesn't mean that your feed is going to be accepted on an LLM that has been, like, one of the biggest, points of contention across the entire industry for the last six months. And then furthermore, they want to say, wait a minute, not only do I need to be live on there, but I want to be Able to complete the ability to, to push a button and make magic happen. [09:33.2] So I'm gonna talk through that as a consumer, right? You go to Perplexity and you say, hey, I wanna buy. A great example is pacsun, that's one of our clients who's helped enable this. But PacSun jeans, so TikTok told me PacSun jeans are the best around. So I'm looking it up, I'm like, hey, this is my size, I need a PacSun pair of pants. This is all the details that I want and then observe the information. [09:51.2] And I wanna be able to purchase the product on Perplexity, right? Cause I don't know about you, but if I get redirected somewhere else, I, I'm probably not buying the product, right? And so that seems like a really simple ask, but it's quite intense. And by intense there's many systems at play, there's many feeds that have to go out, there's a whole backend solution that you don't really have visibility at from like a marketing lens that you then have to work with IT teams and figure out all of the needs as the business to then connect all of the different features together and then also get the approval from the LLM which is like pretty challenging. [10:23.1] And so the exciting part is that we hear these needs. Some people come and say, hey, we just need enriched data and what enrichment looks like to them is filling in missing fields that don't exist in the dot com and they actually need it in order to power one of their marketplaces or for them to power any of their paid ads because they know that they need a better performance lift or it can come all the way. [10:41.4] The full suite of I have this problem. I need to be live on Perplexity and I need to be able to enable the agentic checkout experience. And so for us I get thrilled because we have ears to the ground from a team of product folks and they come in, they say, wait, this is the issue, let's figure out how to build it. [10:57.0] And they go back. And so that's why Commerce is pretty uniquely positioned because we have BigCommerce so we have the E Commerce platform lens on how systems need to integrate and then we also have the Feed lens so that we are able to say, oh, we can move data from one place, send it out, we can make sure that this information is coming back into us. [11:12.6] And the best part I think is that we talked about this earlier is that we have technology with the heart, we have people who are behind the scenes, making sure things are moving in the direction that they need to and if not they are getting notified in real time. So we have these like notifications, these benchmarks that give flags and flags are important because that means our team can see it, respond to it, resolve it, which is so important to a brand like Dell, if you can imagine. [11:35.1] But furthermore, they don't necessarily have the answers on how to solve for issues in a totally new channel that's never been, except for the last year. So I know I'm going off in many different tangents, but I'm trying to kind of zoom out on the complexities of these different asks from a brand by brand basis. [11:50.9] And so what we're able to do is we built out the ability to complete the agentic checkout experience with PacSun and that was incredible. Shirley (Gao) was thrilled, because we were able to create a solution that she had been trying to figure out and we were able to do it, I think in less than 30 days. We were able to stand it up, which is like insane, you know, coming to us. [12:08.3] And it was at Holiday too, which you can imagine Holiday is kind of a crazy time for anything. So we come at Holiday. They asked for this thing, they needed it by a certain deadline and we were able to stand it up. And now you can go to Perplexity. You can find the PacSun jeans, you can get all the information you need, size, you know, and then you could push the button. [12:24.7] It can sync either to your PayPal connection. You can use, you, know, whatever's in your PayPal wallet and then you get your confirmation email from PacSun as if it's that expensive. And so this is thrilling because it's changed the way that we are looking at our expectations as consumers. [12:40.7] Right. We shop on a dot com, historically, now we are shopping on the LLM and we need there to be no friction and we need it to be quick and have it be seamless. Otherwise we're probably not going to go do that again. Right. I think that that was probably answering the question in three different scenarios. [12:58.6] Yeah, well, I mean it's a complex set of things to consider which actually leads into the next thing I wanted to ask. So a lot of the clients that we create content for are small and medium sized businesses who are on commerce and in these cases, like I think they really want to be part of the cutting edge. [13:17.9] They also have more limited teams and resources to get things done. What are the most important things that brands in those situations should be thinking about and Doing? That's a good question. I think the most important thing is having, the ability to understand where your data stands in terms of its And when we say enriched data, like I mentioned earlier, everyone's going to have their own range. [13:39.8] If you're a small and medium business, you still need to make sure that you have the right, details within the product. So that when I am looking at a search query, then they're able to bring in like the FAQs, some of the reviews on the performance, all of the details that's required for there to be search relevancy. When we are searching a long string of keywords, it's not like a caveman, like suitcase. Right. [13:59.3] That's long gone. We talked about the examples, so I think everybody needs to assess their, readiness. And when we say that, we're talking about where is your data at? Do you have data quality? What is your quality? Where does the data live? In one centralized place. And then I think the next step is even if you're a small and medium in business, you should still have interest to be considered on various LLMs. [14:19.7] You should get the outreach to figure out if you can be approved. You should find a feed provider if you're not working with them. Feedonomics is great, I'm a little biased. But you can get with Feedonomics, for example, and they can submit the request on your behalf. Right. You can start to get prepared. But then furthermore, you need to have kind of this internal discussion with your business where you decide, okay, well, where is everything living? [14:39.1] Where are my assets? Right. Where are my reviews? Where is my faq? Where are we having all of the details that make our brand really unique? And how can we get it in one centralized place? Because all of that information is going to be really relevant when they do decide, wait a minute, I have to be on LLM because everybody else is right? [14:54.8] Not so much on like the, you know, there's a need to have AI. What we said earlier, don't do AI for AI sake. But if we know that, I think there's like a pretty crazy number where, you know, this is all going to be the expectation for consumer behavior by 2028, I think is everyone's going to be moving to LLMs as their first point of search. Right. [15:12.6] And if that's the case, even if you're small, medium, or you're the big kahunas in the industry, you have to be present and you certainly have to be visible. And if you're not, you're in trouble. So I would say preparation, assessment, and then understanding where all of your assets live and how you can centralize them is really important. [15:29.5] The way we summarize it in Feedonomics is data enrichment, leveraging that, having standard best practices on your feed. So what is optimization? And then on top of that, like, have one centralized location of a person who you can work with to help navigate as you're trying to get expansion on various channels. Awesome. [15:45.3] Well, thanks so much for your Do you have any final words or hot takes to close with? Oh, hot takes. You know, my hot take that I took away from all of the AI sessions today is that this is, this is the topic, this is the thing that everybody is asking. This is the question that every business is having to discuss with their board, with their senior leadership. [16:02.7] And it's understanding that there's a lot of AI out there and where to start. And I think just asking the question internally and coming up with a roadmap, partnering with people that you trust. Businesses like Commerce, Economics, bigcommerce Makeswift to help kind of explain the vision. [16:18.6] When you explain the vision, it's helpful to bring it into reality. We've seen that some people are going, hey, we've kind of been talking about this behind closed doors. People got on stage over here and said that, and then they said the biggest issue we've ever done is talk about it behind scenes and not get loud so that we can reach out and get the partners help. Because there might be solutions that you're trying to solve for that already live within the current partnerships that you have. [16:38.4] So I guess that's like my hot take. I don't know if that's like the coolest thing I've said, but it is definitely an important hot take. And just in general, I think, continuing to watch this ever evolving scape is really exciting as, Sharon gee, our, SV of product Feedonomics, so she says at Commerce, leading the Feedonomics brand, she always says, you know, like, this is. [16:56.4] Things are getting exciting again, right? And it's been a little bit of a lull for the last 10 years. And we have a really cool opportunity to like, hey, this is like, this is the time to be innovative and to make change. You know, take risks, take chances, learn, trial, find people that you can get. Wait a minute, that was what you did over here. [17:12.9] Look at case studies, see what people that look like your brand and what they're doing, and then, you know, be able to move agile and be swift, and I think it'll be really incredible to see what people do for the next year. Awesome. Well, thanks again. Yeah. Thank you.