#98 Reimagining Grocery Shopping: Instacart's AI-Powered Future with ChatGPT and Artificial Intelligence の全文書き起こし

that you that you stipulate. And so I'm kind of curious if you can share a little bit of kind of the back end and the broader like from a corporate standpoint, you know, what is Instacart? What are the different pieces involved? 05:31 Ali - Yeah, absolutely. And so Instacart is particularly interesting, because it's a four sided marketplace, we're used to multi sided marketplaces and technology, but Instacart is a particularly interesting one, because it has so many of these different stakeholders, basically, in our ecosystem that make it tick and create that magic that you described, like I shop, and then someone pops for me and then delivers the products to my door a little bit later. So we've got, you know, the Instacart platform itself, one side of the marketplace, we've got our retailers, a very important part of the marketplace, so we have 1200 retailers on our platform, which reaches 95% of households in the US and Canada. We've got brands, which is where I spend most of my time. And so brands who are, you know, advertising on the platform and reaching all of those consumers across those different retail partners. And then we have our shoppers who are really what make the entire system tick and are the people that you actually see coming to your door and that you're interacting with and are really the face of Instacart in a lot of ways. And so, you know, this all comes back to the user who is just interacting with all of these different pieces of the marketplace and how we actually are able to then connect all of those together and drive value for each part of the marketplace and then make it work for you, the consumer, as you're placing and receiving your order. And one of the things that I really like about, you know, thinking about consumers on the Instacart platform, retailers, brands, and shoppers is all the incentives are very much aligned. If you zoom out, everyone wants to sell stuff. Everyone wants to create more orders on the marketplace. So brands want to sell their products. They want to reach consumers and sell products. Consumers are coming onto the platform looking to buy something. They have very high intent to buy something that day. They may be open to discovering new products and we'll get into that in a moment. But they have this high intent coming to the platform looking to complete a transaction. Retailers want to drive more sales. Shoppers want to get orders and make sure that they're able to, you know, shop for consumers and complete the purchase cycle on their end. And so it's really exciting to think about this from an ads perspective then. So when my team is thinking about how to drive maximum value across the marketplace, those incentives are really quite aligned. It's a world where the ads can truly be part of what is actually making the experience real. It's not something that we need to balance in terms of, oh, we might like be interrupting a user or we need to consider, you know, how are we bringing their

attention away from what they might be otherwise doing on the platform. It's all in the service of actually helping to discover and buy products. 08:04 Jessi - Yeah, that's super interesting. I hadn't thought about how like, I mean, it makes sense, but just the intent to buy is so much higher than other advertising areas. Like you said, you're not interrupting someone. The person is literally like you're literally trying to add things to your cart at the moment. And I think, I mean, the like the algorithms behind it, I've always I've been impressed from the beginning of like the ability to track of like, hey, you usually buy this every week, you didn't add this this time. And I'm like, oh my gosh, you're right. I forgot, you know, and it doesn't even though it may be, it may be sponsored sometimes, but it doesn't feel like that as a consumer, because you're like, oh, no, I totally needed that right now. And so I think it's it's such an interesting. Yeah, it's just I think it's such a fascinating platform. So that's really interesting to hear the four sided piece. And like when you say like 1200 retailers, that means like 10s of 1000s of actual retail doors, right? 08:57 Ali - Like exactly, exactly. So each retail banner, so to speak, has many, many stores, depending on the size could be many, many 1000s could be a few. But we have a wide variety of retailers on the marketplace. Exactly. 09:07 Jessi  - Yeah, okay, cool. So yeah, we're definitely going to keep digging into, you know, AI and some of the new features. And I've been playing with some of them. So I'm excited to talk about them. But like, can you kind of talk us through like how like ads have evolved on the platform, particularly, you know, from the brand set, like maybe, you know, the brand experience to have how that's evolved of what's been available for brands, like types of ads that they've been able to serve, and then how, you know, when and how AI has kind of been entering the picture into this advertising piece, like, I'm curious, you know, to kind of get us started in that area. 09:41 Ali - Definitely. So we've been on quite a journey over the last couple years, really, on, you know, driving continued improvements in the core ads experience, and then creating many, many more ways in which brands can reach and inspire consumers. And so really, we started with sponsored products, you talked about sponsored products earlier, and they're exactly what they sound like. They're specific products that can be promoted across the stages of the shopping journey. The nice thing about our sponsored products offering is that it's not just a search ads product, it can actually be promoted products throughout, you know, browsing the store front of a retailer on your way to checkout when you're considering other items you want to add to your basket. But they're really anchored on specific products that can then be promoted that the advertiser is choosing to bid for basically. And so that's

a, you know, relatively standard offering across the retail media space. I think what really sets us apart there is our ability to adapt across the multiple retailers that we offer on the marketplace and help to really tailor to the preferences and likelihood of interacting that a consumer may have. And then finally, in the results. And so this is a new piece of what we've really offered in the sponsored product space over the last year or so, which is not just looking at attributed metrics, not just looking at, you know, return on ad spend, all those great bread and butter metrics that anyone advertising is familiar with, but actually getting to the heart of what is the incremental impact of sales driven by ads on Instacart. And so we've been investing in sales lift metrics, A B testing, optimized bidding, which I can talk about a little bit more and kind of the ML and then AI space, but really driving the most efficient results we can measuring that clearly, and then demonstrating to advertisers what true value and incremental value that platform is able to drive. And so that's sort of been our journey of, you know, offering a core sponsored products offering, but really focusing on how we can drive and demonstrate the value of that offering to advertisers in a way that, you know, really gets at the heart of what is truly going to be a major contributor to an advertisers or a brand's business. So that's really the sponsored product space. And then we start to get into all the beautiful new formats, the more inspiration focused and discovery focused formats. And so those have also been new in the last year or so. And so this is all about expanding a little bit of funnel. So we talked earlier about Instacart's high intent, consumers come in with an intent to shop, they may have a list in mind, they may have that list written down somewhere, we'll get to where AI can then come in and actually translate that into an easy shopping journey. But a lot of times consumers do have something in mind already, but what is the fun part of grocery shopping? The fun part of grocery shopping is you walk into the store, you look around, you get inspired by things, you decide to try something new. And we really want to create that feeling on the Instacart platform. If it's just a transactional platform, it's a little boring, like it may not be as fun, you may not have that opportunity to discover something new. And so we focused on kind of finding that sweet spot between Instacart's ability to drive that transaction and close the sale and adding a little bit more inspiration and storytelling so brands can reach consumers in a kind of more exciting way. And so we've launched our shoppable video and shoppable display products that are specifically designed to do just that. So pair of rich media creative with a carousel of shoppable items that can help to then drive those kind of novel additions to a card and drive discovery, especially for emerging brands who are kind of maybe earlier in telling their story and trying to get that excitement level to build a new loyal customer interaction. And so that's been another big piece that we've invested in over the past year to really start to create a bit more of that fun and inspiration on the platform. 13:28 Jessi - Yeah, interesting. And for some of that content, from a brand side, are brands generally logging into an Instacart backend to add those kind of videos and those kind of things? Are they using I know I've used some third party tools even a couple years ago that were sometimes AI powered for some ad bidding, which we can talk about some of that too. But I'm curious from a backend perspective, how are brands getting this content onto the platform? Do they have the opportunity to update their listings or does some of the data just pull from the retailer information? How does the brand's data end up there to then be used in these pieces? 14:07 Ali - Yeah, so we have a full fledged ads manager offering. And so this is similar to what you'd see on ads platforms that provide that sort of self-service tool that you can log into, create campaigns, edit campaigns, edit your creative, monitor performance, all of those wonderful things. And so that's a fully self-service tool and that's available to any advertiser who's advertising on Instacart. And this is actually one of the first things I focused on as a product manager in ads was this sort of advertiser facing experience. And it's always an area that's just quite close to my heart for that reason, making sure we're really clear on how we're actually organizing our product portfolio. We just launched goal-based or objective-based buying in our ads manager that helps to take all of those different products I just mentioned, the different formats, different bidding options, and actually buckets them into goals that advertisers can use to understand what to put together to achieve reach, engagement, and sales on Instacart and then the metrics that go along with that. And so that's an area we've been investing in quite a bit. And then to your point, there are also API implementations. And so we manage a read and write API that our API partners are able to use as well, which is really great when you're looking at this rise of retail media and many folks are looking for an extra level of expertise or management or optimization that can be offered by those API partners. They're also able to basically tap into our system via our API and then manage on behalf of customers or provide those additional services that they are so great at doing. So we have a really strong relationship with our API partners as well. And it's been something that's been really great to continue to grow, especially again, in this context of retail media growing so quickly and supporting those experts in the space has been a really important part of the whole ads ecosystem. And then the last thing I'll add is when you're advertising on

Instacart, you are already a brand who has sales of your product represented on Instacart through the retailers that we work with. And so if a product is available in a given retailer, when we are ingesting that catalog and supporting shopping and browsing of those products on Instacart, that's sort of the first step in the process. And so you're on the physical shelf, then you're on the digital shelf. But then what ads offer is then this ability to promote to a level that may not be possible on the physical shelf or may be much more dynamic and personalized and performant. And so especially again, coming back to emerging brands, one of the things that's been very exciting to me is that an emerging brand may not be able to have the level of shelf space or promotion inside a store that an established brand might, but the physical shelf, because the physical shelf can't adapt, it can't say like, oh, you know, I think that you may be really be interested in liquid death water, for an example. But on the digital shelf, that can be completely personalized, it can be dynamic. And so we can offer these abilities to connect with consumers in a really visual way, in a really prominent way, and then drive additional sales and discovery, thanks to that level of customization and personalization. 17:10 Jessi - Yeah, that's a really interesting point, because yeah, you think about how, like you said, as an emerging brand, you often don't get the greatest placement. You might be in the top upper corner of a shelf, you might be down on the bottom. And if you're just standing there looking at the shelf, you might not notice that brand. But if you're at home, and you see it and you click, you know, an experienced shopper is just going to go and they're going to be looking directly for that brand, and they're going to grab it. So it totally, it totally changes the game there on kind of leveling the playing field a little bit out for emerging brands to, you know, at least digitally, you can you can be next to everybody else and be top of mind. 17:48 Ali - So that's exactly Yeah, it's so funny, because you know, when I get unsolicited feedback, sometimes from my friends who are shopping on Instacart and actually have discovered new brands that way through, you know, a shoppable display unit or something is like, Oh, I saw this thing, I bought it because of the ad, it was really exciting to never knew about this brand before. And I mean, that's the kind of unsolicited feedback about ads that I love to receive. And so I love the fact that that feels really authentically part of the experiences as part of this platform. 18:13 Jessi - Yeah, awesome. Um, well, there's, you know, there's a lot of different AI pieces in here, but I'm wondering if you can talk a little bit about the partnership specifically with chat GPT, like, what are some of the features that are powered by chat GPT specifically? And like, you know, what do you, is there a roadmap as part of that, or for things that you see down the road, and maybe that's too far ahead, but kind of curious about the chat GPT piece specifically, and what that partnership is kind of looking at, and what's the best way to do that. And what that partnership has kind of looked like to explore and start to implement. 18:46 Ali - Yeah, so this has been a really exciting part of the Instacart journey. I feel so excited that we are early in this technology, we're leaning in really hard, we're getting excited and connected with partners in the space. And it's moving so quickly that even we talked about JJ, our, our leader in this space at Instacart. And you know, he is an absolute expert in this space. And even he says he can barely keep up with the level of change. And it's, it's just such an exciting time. So on the chat GPT front, we basically have two directions that this has gone in. So one is bringing the power of Instacart into chat GPT. And this is more of the plugin angle. And so wherever we're able to help people complete a transaction on Instacart, by plugging in our capabilities to where they are using chat GPT, that's one angle where you can basically just make that, you know, we talked about pixels into atoms, we can make that actually come to life in the context of a chat GPT experience. So that's one angle. The piece where it's really starting to come to life and evolving on Instacart itself is with our Ask Instacart launch, which was just recently announced. And so this is all about bringing natural language and context and exploration and you know, much richer queries and questions and answers that can actually be asked on the Instacart platform. And so when you think about how people have, you know, been trained to search, if I'm thinking about, oh, I want to, you know, plan for a barbecue this weekend, like, okay, well, what does that mean? I need like the meat, I need maybe the tools, I need the, you know, the whatever it is, like the charcoal, like all the different pieces that are going to create the barbecue experience that I want. And if you can just simplify that into what do I need for a barbecue? And then I actually get a set of products that are related to that or what's for dinner, which is kind of like our perfect go to example of a really complicated question that people struggle with literally every day. I know I do. The idea of actually taking in that question and understanding it and then recommending products as a result that you can then purchase and buy and get delivered to your door a few hours later, like that is a super magical experience. And just the very beginning of what this technology can do. So I am really, really excited about this direction. And the fact that we've been able to do this so early, we have so many rich opportunities and so much enthusiasm inside the company around shaping additional experiences like this is really, really cool. I am particularly excited about not just, you know, the question and answer

and like the chat bot kind of interaction, but also just how this threads through the entire experience. The idea of creating just a more personalized, truly dynamic storefront experience when I'm browsing what's, you know, at my favorite retailer, it can just be so much richer and so much more responsive thanks to this technology. And it's just so exciting to see where it's all going to go. 21:43 Jessi Yeah. Oh, yeah, I definitely I want to dig a little bit more into both of those areas. The first one on the chat GPT plugin side, because when I from the moment I saw the like TED talk, I was like, I want to play with that, like I was. And so I have been playing with the plugin. And it's been super fun, and just like, really useful, like for, you know, meal planning, like I'll put in I've like, tried to, you know, like, you know, trick a little like, see, like, how deep does the knowledge go, like saying, like, okay, I'm planning a meal for four people, one has celiac disease, one has a shellfish allergy. And I'm like, Okay, does it understand that celiac disease means gluten free, you know, like those kind of things. And like it did, and it popped out like a really great meal plan. And I was like, Okay, that looks good. And it was like, Do you want to send this to your Instacart? And I was like, Yes. And then it just automatically goes, you know, it's just in and I was like, this is awesome. Like that saves so much time. I am not I do not I'm not someone that really enjoys a lot of the meal planning process. I was like, this is a game changer for me, like to be able to like interact with a bot essentially to help me make decisions that I don't want to have 22:50 Ali to make. I was like, this is super helpful. So yeah, and this is the amazing thing is like, you know, that's not needing to now be individual training data, and everyone's creating their own models and all of this. It's just it's technology that we can share, and actually make truly useful and connect across all of our interactions, it's going to be incredibly exciting. And that that reaction of like, Oh, can I trick it? Oh, actually, that was pretty good. I feel like that's that's also my reaction when I start to play with these things, like, well, it understand and following up on a previous question. Yeah, it does. But it's Yeah, it's truly my darling, just how quickly 23:24 Jessi this is all moved. I'm curious, the my first question, especially from like, you know, from the start of CBG angle, when I was using it is, you know, I think there was something, it said I should make a you know, it suggested making a gluten free cake. And it suggested, you know, it wanted me to make it from scratch. So I was including the ingredients and like one of you know, a couple items were branded like a branded cocoa powder branded and I was curious, like how you see the branded part like evolving? Like, how does it choose a brand? How can like an emerging? How do you think an emerging brand will

able to be like able to be pulled into some of these like searches? Do you think there'll be opportunities for sponsorships? Will they be able to like update their listing to be like, you know, rich with data for the AI to pull in? Like, I'm curious about the brand choosing process and I'm sure it's evolving. But you know, I'd love your perspective. 24:17 Ali Yeah, definitely evolving. And I think as you're seeing across the industry now to everyone's trying to figure out like how in this completely new environment, how do ads and brands and advertisers kind of play within this space. And so we are really taking that exploratory approach as well. And I definitely I don't want to promise yet like what that solution is going to be. But there are a few things that we already know will work and are working. So when you receive, you know, when you're doing when you're working with us, it's Instacart and you're asking it a question that will ultimately create a set of products that are recommended that you buy that are based on the question that you asked. And so if you continue to navigate and like look at the products that are that are returned as a result of that question, you can see sponsored products as part of that experience. So that's kind of like the basic first stage is, you know, wherever we show products and sponsored products are a good match, we will also be able to return sponsored products. So that's kind of by doing nothing on the advertising side, this is still a relevant query essentially that the ad can be returned for. But there's so much more beyond that that we can do. And so as you might imagine, we're, you know, we hear a lot of enthusiasm from advertisers about this. And exactly as you said, you know, it's a great way to encourage more discovery of both emerging brands and then products that folks may not have thought of buying on Instacart. We talk a lot about how it's not just fresh grocery, it's also your household items, your personal care items, like everything you need to kind of complete the totality of the task that you have at hand and that the Ask Instacart experience can help guide you toward. And so we are definitely going to be exploring and iterating on kind of richer ads, ads experiences and sponsored experiences, and ways in which brands can get involved in in being a part of this experience. So realize I don't have like a lot of specifics on exactly what we're going to do yet there, but that's for a reason because we really want to see where this evolves, how consumers are interacting, and then what other experiences might really make sense to drive new and exciting value for brands and also for consumers who are just learning how to navigate and adapt to these new experiences. So there's going to be a lot up ahead for us, but just starting on those basic terms of making sure that sponsored products are part of the experiences is really the place that we've started and will continue to grow from there. 26:32 Jessi - Yeah, yeah, no, that's, it's really interesting to just hear how you're thinking about it evolving and because my first my, you know, among the many thoughts I've had of trying this, I'm like, this is, this would be so interesting as an opportunity to like, you know, train, like train the app on like also like your preferences kind of thing. Like, okay, like, you know, maybe it knows that when I, when I say I want to make dinner for friends, I also want to choose climate friendly brands when I can, or I want to choose women or LGBTQ+, like own brands when I can. And so that it could like help prioritize based on your like, because we've, we've had like social nature on our show, and they talk a lot about like how like, you know, wellness consumers are, you know, very driven by their different, you know, the different areas that they feel passionate about. So be able to like, maybe even bring those into the shopping experience. I was like, that'd be cool. For me, from a consumer, I was like, if I could discover brands that match my values, like right in using the AI using Instacart, and then it all shows up in my door, that's where I'm kind of thinking, like, that would be so cool. 27:33 Ali - Yeah, I love that. And I think this is also kind of speaking to the power of the Instacart data specifically, in enhancing what you can do with technology like this. So for example, the power of our catalog, our catalog, again, spans all those different retail partners that we that we work with, and all the US and Canada. And so we have a really rich set of first party data. And we have different attributes on different types of products. And to your point, you know, it could be gluten free, like gluten free items, it could be kosher items, it could be taking that next step ahead, the idea of like eco friendly, or, you know, socially conscious brands, or whatever it may be, the idea of being able to take in to build out more of these attributes, take in more of those attributes as data, and then be able to return a personal, a truly personalized set of recommendations for the consumer. That's really the promise of this technology. And I think even starting where we are now with the richness of our catalog, it provides a ton of different experiences and options and personalization that, you know, build on what Instacart has been doing for a long time with, you know, ML powered personalization, and remembering what you've bought before and providing you with suggestions that are going to resonate with you. But it's just that much richer and that much more resonant to like build these authentic discovery moments with the consumer. So that, yeah, absolutely. I think that's, that's just a great vision for the future that we're all moving toward. 29:00 Jessi - Yeah, no, that's awesome. And I, I do want to dig more into the Ask Instacart too. But like, also this, this makes me think of from the brand side, is there any like, best practices or like,

especially for an emerging brand, say you log into the Instacart, you know, portal, or you're logging in through a third party, you know, that's accessing the API. Like, is there anything that brands can do, like as far as like adding more data to their listing or adding more keywords, or things that you see really help an emerging brand set up for success? Because I know sometimes like, say I, I might be searching for a brand that's smaller, I know about them because I've accessed them and you know, their listing is, you know, there's maybe no image yet, and they may be new at the retailer, but you know, like, it's very basic because they're just kind of getting set up. Is there anything that you see emerging brands that really do well on the platform set up on the backend really intentionally the other brands can maybe take away? 29:55 Ali - That's a great question. And I think that really does actually stay in this theme of the power of automation. And so, you know, maybe some of that's good old fashioned machine learning and not yet AI, of course, it's all getting there. And so we've really invested in creating a pretty, you know, I don't want to say like, like 100% like ready to go one click setup, but pretty easy to use and set up platform where you don't need to think about each individual keyword you want to add, we are able to infer that from the products that you sell. And we are able to optimize toward an outcome that you would like. So if you want to maximize sales, you can use our optimized bidding capability, which will automatically manage bids on your behalf. So you don't need to think about the bid value for each keyword and all of the kind of, you know, nuts and bolts of pay per click advertising. And so if you really combine that ability to figure out which keywords make sense, and then drive toward the outcome that you would like, that's a that's a relatively easy way to get started for an emerging brand is something that I've been really happy to see us be able to focus on and launch with great results to make sure that we're, we're not adding too much burden to really set up a campaign that's going to perform really quite well. And so that's, that's really what I would recommend, especially as we're thinking about these more dynamic and personalized experiences that we've been talking about, that automation is going to do the right thing in those experiences, you don't need to think about, oh, like, what questions might people be asking and ask Instacart that I want to show up for, that's all going to map back to the relevant keyword, the relevant item. And so it can all quote unquote, just work, which is really exciting in a space like this, where additional, you know, manual twiddling of keywords and bids is probably not going to be the way to take advantage of those new consumer behaviors. So that's, that's something that I expect we're going to rely on even more as we continue to move in this direction. And then as you

mentioned on figuring out, you know, the images and making sure products are represented correctly, that's another piece of what we offer in our self-service tool. So we have a product library feature where brands can come in and figure out, you know, what images might they want to suggest instead of the default image that we get from our catalog or from the retailer. And then how else do they want to make sure that their brands are properly represented? And so this is, this is a great way to kind of bring those two parts of Instacart together, like how your products are represented out of the box, and then how you actually want to promote and manage your campaigns and monitor their performance of them. 32:22 Jessi - Oh, great. Yeah, that's, that's super helpful, all super helpful information. And so on the on the Ask Instacart piece that you were referencing, and this is something also that like I was playing with in the platform. And like, you know, in my own interface, you know, it popped up with like, here's what people are wondering that you can ask Instacart. And it was like, for me, what popped up, and I'm curious about how this is catered per user, it was like, what are the most popular flavors of Spindrift? And I was like, this is absolutely a relevant question for me, because I drink so much Spindrift. But I was just like, I was like, oh, this is fun. And so I was kind of playing with the different prompts that it suggested for me. And so I'm curious, like how, you know, personalized to the user is that how like, when it's defining popular, is that based on like, you know, shopper data within the platform? Like, how, you know, how does that work? How is it helping generate some of those prompts per users? Like, I'm curious about that. 33:13 Ali - Yeah, definitely. And this will also, I mean, of course, continue to evolve over time too, as we continue on this whole journey. So in general, across Instacart, you know, we've been using ML and increasingly generative AI to personalize what we show to consumers. And, you know, the, probably stating the extremely obvious, but as many, many folks who have worked in this space know, you can do a really good job curating something manually, or you think you have an idea of how products should show to a given user or set of users, and machine learning and increasingly generative AI will outperform it almost all the time. And so we are always working on figuring out what are the right sorts of, you know, signals and personalization or personalized experiences that will best resonate with the consumer, whether it's, you know, just the good old fashioned, again, storefront and set of carousels that you're seeing with recommended products, or increasingly now in the generative AI space, where we can actually offer these more dynamically tailored and personalized prompts to help kind of inspire the types of questions that people might be interested in asking. And so absolutely, this will continue to

be personalized based on what we see people doing on the platform, what people like you might be doing on the platform, what you specifically have done on the platform before. And again, when you think again about like how this breadth of in depth of data that we have around how consumers are shopping, and how this is kind of changing based on where they are, who they are, how they're shopping, all of these different things, we're able to do this at a scale across our retail partners that is really quite different from what you would see if you looked at a specific retailer or a specific location only. And so it's really building on that core strength, I think, of Instacart of being able to personalize across all of these different dimensions and then build a great experience, especially now in the Ask Instacart space, as consumers are getting used to asking questions like this, I think it's particularly valuable to make sure that we're giving them a starting point because it's not going to be the same old search for bananas that so many people do. It's going to be like training people to think a little more broadly and expect a different type of response and be excited to explore. So I love that the suggestion for you is so personalized. That's perfect. And that's very much what we'll want to be doing. 35:25 Jessi Yeah. Oh, interesting. And so I'm also curious on the like, it seems like with like, um, machine learning ML that you mentioned, and like the generative AI, these pieces, like it's all evolving very quickly from like, from the Instacart side, like, what does that look like for your team? Like, is it like, is a lot of the updates kind of constantly happening? Is there also updates where like, you know, you're pushing an update, like to the Instacart app, like, you know, how, how long in between kind of big updates, like, what does this look like on the back end for your team to manage? Because it seems like it seems like it's going so fast. And I'm wondering how you keep up and like I said, push updates to people's phones, update the interface, like, you know, what does that look like on the back end? 36:06 Ali - Yeah, absolutely. I mean, it's a very fast moving company. And I, it was something that really struck me when I joined things that I might have expected, you know, might go out in the next quarter. So it's like, okay, it's two weeks good. Like, I mean, you know, sometimes, of course, things could take longer, maybe be more complicated. But I think my, my internal rhythm for what I expected in terms of product releases and iterations has like, massively shortened. And so it is really quite exciting, especially in this space where we, you know, we want to experiment, we want to understand how consumers are responding, we want to try something quickly, we want to even do internal demos on like, what cool thing might we be able to do with this new technology that is happening on a very ongoing basis. And so really, the only limiting factor is, you know, someone installing the updated version of the app on their phone, for example. And so, you know, changes are being pushed out constantly. It's really a, you know, daily and weekly cadence for teams. And so it's a very, very fast pace of change. And I think in a lot of ways, especially in the, you know, the advertising space, that doesn't mean we should be constantly changing things every day. It means we can fix things every day, we could try new experiments every day, but we need to make sure that we're building a narrative that actually sits well with advertisers and that brands can kind of understand and respond to and figure out kind of where our heads are at in terms of how we want to actually evolve the platform. So finding that right balance of how to actually build like, a B2B narrative, for example, versus this fast pace of iteration, that's something that also takes quite a bit of balance, I think, internally. And that's something that I personally really love, like the idea of taking all of these changes in the technology and then figuring out, okay, how does this start to make sense as a cohesive strategy and as a narrative and set of capabilities that we can now start to frame a little bit more clearly for a business to really understand what's going to drive their bottom line. And so then more generally getting back to your question on generative AI and how quickly that part is moving. So the most fun thing for me has been to see the engineers on the team just absolutely go to town on demos, trying new things, figuring out different applications, whether it's literally, you know, easier debugging of data, like internal productivity tools to ways in which we might be able to help advertisers better understand what to do within our ads manager interface, things that are more advertiser facing, ways we can help our sales and support teams to better understand how to answer questions for our customers. And then moving into how do we improve the core experience and create more relevant personalized moments like we mentioned throughout the entirety of the core Instacart app experience. And then finally, more of like the chatbot type like interactions, which of course have, you know, captured the hearts and minds of so many people around the world. And so all of these pieces can really, you know, be something that engineers just get really excited about pursuing and start to build momentum around. And so my approach to this personally has really been, I want to make sure we're asking the question, like, what is this new technology actually enabling us to think about or do in a way that was totally unimaginable, like six months ago? And how do we make sure that we're really thinking about and like pushing ourselves forward on the type of product development that we want to focus on? And let's also just try some cool stuff and figure out, you know, time for demos and time for smaller pilots and things where we

can just test and learn. Because in a space like this, where you just have no idea what the next two months are going to look like, we need to keep that level of like iteration and excitement. And I love seeing that really just bubble over on the teams. And, you know, they can just dig into so many exciting things without really any prompting at all. 39:49 Jessi - Yeah. Oh, that's so cool. It sounds really fun. But yeah, it does. It is interesting to balance because like, even if the technology keeps progressing, like that doesn't mean that the consumer is ready for all of that in there. Exactly. Immediately. Like you said, like ask Instacart, you have to kind of learn how it works. It needs to give you some ideas so you can start to learn how to use it. Or from a brand perspective, if like the way that you buy your ads or everything in like your backend, if that all just like, you know, changes too quickly, like, you know, you have to help people along the change curve. And so it's both exciting, but also like, you know, you have to you have to manage all those pieces. 40:21 Ali - So true. Yeah. I mean, at the end of the day, people are people and they still need to learn and get comfortable with new things. And so that's, that's really our responsibility and like shaping something that actually resonates with real human beings, real businesses, real brands, and so on. 40:33 Jessi - Yeah. Awesome. And then I want to circle back a little bit to the, like the ad interface that brands have access to. And because I believe there's some like, you know, some automation that can happen where you can kind of like set bids and those kinds of things. And then there are, like I said, you know, and we talked about there are third party companies, like I've used one where it, you know, it's, it will kind of do the kind of set bidding for you and kind of set it, forget it. You set up a bunch of rules, those kind of things. I wonder if you can talk a little bit about some of the opportunities within there to kind of set some automation as a brand and set goals. You mentioned some of like, you know, kind of maximizing sales, but I just want to leave some space to talk a little bit more about the opportunities to use automation to set up a campaign with an Instacart that really fits your emerging brand. 41:20 Ali - Definitely. And yeah, this has been a really exciting part of what we focused on, again, really over the past 12 to 18 months. And so, you know, kind of moving from a world in which, like I mentioned before, you can imagine an interface where you're specifically describing, you know, which keywords you want your products to show up for, which bids do you want for each of those keywords, and many other ways in which you could kind of manually identify how you want to actually run your campaign. But when we're able to have so much context and so much first-party data around how consumers are responding to products and how those products fit into

our catalog and whether they're likely to be in stock and all of these different things that we're able to do from an Instacart perspective, it helps us to really set up a pretty darn good default kind of setup for a campaign without much work at all. And so I spoke earlier about how, you know, a brand can simply, once they start to advertise with us, they can decide which products they want to promote as, you know, a set of sponsored products, and then we can figure out which keywords those are likely to best correspond with and what's going to perform. And then with our optimized bidding offering, we can take that work of selecting the specific cost per click bids off the advertiser's hands, and we can actually just optimize to drive the most sales possible at the most efficient level. And then the advertiser can optionally set, you know, a return on ad spend guardrail or whatever they may choose. And so when you think about really that basic setup of drive sales on Instacart with ads, that is our most performant offering is just utilizing optimized bidding. We've seen with A-B testing that optimized bidding consistently outperforms manual bidding and drives more sales at more efficient prices. And so it's been really rewarding actually to see just how well that's been embraced by advertisers, especially by emerging brands who, for obvious reasons, you know, are just working often with way fewer people and they need a little maybe a little bit more help or a little bit more understanding about how to actually drive that maximization of sales. And so setting, like essentially setting and forgetting a campaign like that is a really great way to just take, you know, get the most efficiency possible and the best results possible with Instacart ads without all that much work. And so then what I think is particularly exciting for me is like the ability then to go back to that first party measurement and the strength of how we're able to actually measure the overall effectiveness of those campaigns. When you look at A-B testing and you look at SalesLift and ways to actually prove the bottom line sales impact of our sponsored products campaigns, that's consistently coming back with really strong results. And we've seen these, you know, double digit SalesLift results really consistently across our platform. And that is really utilizing the best of this automation and technology that we're able to continue to scale and improve over time. So you kind of, you know, cross that with the ability for generative AI to do even more personalized drive, even more engagement and sales coming from consumers who are really interested and engaged in deciding to buy something based on an ad, that will only get better. And so that to me is really, really exciting that we've been able to scale those results and continue to really invest in both driving the value and then measuring that value for advertisers. In addition to that, you know, kind of getting back to the set

of goals and objectives that advertisers can optimize toward when it comes to driving engagement and driving reach on the platform, that's really where the power of the format itself is coming into play and the ability to, like I mentioned earlier, pair that kind of rich media imagery, the video that shows how the product can be enjoyed or used with the ability to directly buy it immediately with the shoppable carousel just beneath. That's the way in which we can, you know, utilize these objectives or goals to basically easily suggest to a brand who's just getting started that these might be the right products for them paired with specific targeting options, and then paired with the metrics that can actually show that outcome. And so these are areas that will, of course, continue to invest in over time. And we're always going to be working on ways to make optimization better and easier and more easily measured. But it's been a really great journey to start to like put these solutions together for advertisers so they start to make sense and then of course can drive the value that they can really see on the other side. 45:45 Jessi - Yeah, that's super helpful. And yeah, I think I was impressed even a couple years ago and so much has gone into the platform, you know, since then just of how friendly it was for an emerging brand. So a lot of advertising platforms are built for big advertisers because, you know, that's where a lot of the dollars has been. And so I, you know, like I said, even a couple years ago, I was hesitant, like, is this going to be, is this going to be useful as an emerging brand? Is this going to be like built for me? And like it started working very quickly. And and so I like I was super impressed. Like this is not sponsored. I just generally genuinely think it like it was super fascinating to see how like, you know, useful it could be as an emerging brand to start to get data and then also as a tool in your toolbox for retailer partnerships, because so many retailers on Instacart, you know, say like I'm new seasons market is here local to me, it's on Instacart. If your brand emerging brand gets into in, you know, into new seasons, and you're running ads on Instacart, you know, you can share that data then back with new seasons and say, Hey, look how I supported, you know, our product in your store with our sponsored ads on Instacart. And I think it's another interesting tool to build stronger retailer partnerships that emerging brands have in their tool pocket as well with that has really great data behind it. So I think it's all just super exciting. 47:05 Ali - Oh, that's awesome. Well, I'm so glad you even thought that, you know, a couple years ago, because we've really we've come a long way even since then. And I, but I think this idea of really building a, you know, I don't want to tutor on horn, but like really a first class ads platform that is kind of utilizing a lot of what many of us on the team have learned working, you know, across the industry and seeing what's worked, seeing what hasn't seeing how we can best strike that balance of control and automation and which sorts of metrics we really want to lean into and differentiate ourselves versus others in the space. It's been it's been a really, really fun journey. And I think to your point about making sure, you know, brands are really taking advantage of where they have that distribution, the idea of Instacart of buying ads easily across the entirety of the marketplace, wherever those products are sold, as a brand is building their reputation and their distribution, the ads will automatically show wherever they are relevant. And so wherever those products are in stock, and wherever customers are excited to discover them. And so that's another great way to think of it really as a one stop shop across the retailers that we work with, and then, you know, help a brand expand their reach and expand their engagement pretty easily without thinking about individually taking those steps each time they win it, they win the new a new offering of shelf space. Right? 48:20 Jessi - Yeah, yeah. And that's really good point on the, you know, for emerging brands, we don't have nationwide distribution, you don't want to serve ads to people that can't buy your product in other platforms and Instacart platform, it's going to serve it to people that have access to your product at a store that there's a shopper can get to. So yeah, that that eliminates, you know, it's not going to people who can't access your product, which is really interesting. 48:44 Ali - Yeah, I think that's one of the biggest pieces when again, just as we talked about the very beginning, like the idea that you're actually advertising, and then that product will be in the consumer's hands within an hour, like that is a really magical thing. But inherent in that is our ability to understand what's likely to be in stock, what has been seen recently on the shelves by shoppers, where is the product in our catalog, and which you know, flavors or variations are available, all of that is just already true. And then the the advertiser just needs to activate the additional boost of awareness and sales that they're able to get from ad solutions. So I do think that that's a really, it's what makes it such an interesting ads platform to me, just the basic concept of what we're able to do, essentially out of the box, just based on Instacart model is already really interesting. And then on top of that, if we're able to drive, you know, great automation and results and continue to kind of build these core capabilities for advertisers, it just makes it that much better. 49:43 Jessi - Yeah, I'm also curious, one last area I want to check into is like how because you mentioned customer service and how you know, how AI and automation are feeding into that area as well, because I think the customer service piece is interesting on a brand side, because say a

shopper, you know, bring something back to you, and maybe it's expired, or maybe the seal is broken or something, and a customer is going to go in interact with the Instacart customer service team. And that experience might, you know, will have an impact on probably the person's view of that brand, potentially. So I'm curious how like how you use that to and like, I will be the first to say Instacart customer service has always been incredible. So I'm curious how you use technology to power that customer service and keep the experience, you know, super positive for the customer. 50:33 Ali - Yeah, that's it's a great point. And I'm glad your experience has been positive. I will definitely share that with my with my ops friends. I so this is definitely an area where I think AI has a ton of potential. I have personally found that in my own experiences of actually starting to use chatbots like out in the wild and actually get my questions answered, they're pretty darn good. I mean, you know, newsflash like generative AI is great. But it is a really powerful tool in the toolkit. And I think this is something that the, you know, the customer facing teams are absolutely looking into. I think the important piece here is, you know, at the end of the day, like I said, just as we need to get comfortable with new ways of interacting, all of our customers do too. And that human interaction is a really important part of great customer service and the image of a company. And to your point, you know, not just representing Instacart itself, but the retailer and the brand and everybody who's kind of involved in the actual creation of that moment in that order for the consumer. So for me, it's really about looking again for these ways in which the technology can enhance or create a faster experience or a smarter experience. You don't have that feeling of like, Oh, I have to like come through a bunch of help docs and like figure out what I want to do. And then finally, I guess I'll contact someone. If we can answer a question really quickly and easily amazing. There may be cases where that's not true or where someone just simply would rather speak to a human. And so these are all things that we need to balance as we figure out how best to utilize this technology. And I think it again comes back to really first figuring out what can this enable us to do? And then what should we do based on it? And that's going to take a lot of iteration, a lot of exploration, a lot, a lot of really smart people, you know, thinking through the implications of, of this, not just, you know, not just for Instacart, but across all of our interactions on online and increasingly offline. So I went a little broad there, but I think generally speaking, this is definitely an area that we're digging into. And I expect a lot of companies similarly to start looking at how generative AI can not necessarily replace, but enhance the experience of needing help, finding a resolution and finding answers to your questions. And, you know, same thing goes for supporting brands, supporting advertisers and making sure that we have a really excellent, you know, sales team, which we do. And they're, they're absolutely wonderful and super knowledgeable to work with. But if we can help to quickly answer a question that's on someone's mind without needing to go through that whole process, all the better. 52:59 Jessi - Yeah. Oh, great. Well, this is awesome. Is there anything else that you wanted to share or leave us with, or I'll make sure in the show notes to include, you know, links to Instacart links to the brand interface for Instacart, but like, is there anything else you would like me to include in the show notes or you want to mention or shout out that we didn't cover? 53:14 Ali - I think for me that it's really just this excitement about starting to mature and scale as an ads platform and having built out a lot of solutions over the past year or two that have really enhanced the core offering that we can help advertisers use to drive value. And it really comes back to driving value and then understanding what that value actually is. I think that that is really what's always top of mind for me, like working in this business and working specifically on ads at Instacart. And now having done that and having kind of reached this next stage of maturity for our platform, along comes this technology that is going completely to transform so much of what we do and enhance so much of what we've already invested in. So for me, it's both, you know, reflecting on like what we've been able to accomplish as an ads platform, but then now, wow, look at the future. Like the future is here. There is so much more that we're going to be able to do. So for me, that's when I think about AI and then, you know, where it's going to help us go as an ads team and as an Instacart team. That's what's really exciting to me. 54:15 Jessi - Yeah. Oh, awesome. Yeah. I'm super excited as well. And just so appreciate you coming and sharing with us today. This has been super enlightening and helpful. I think our listeners are going to really enjoy it. So thank you so much for spending time with us and sharing. I really, really appreciate it. 54:28 Ali - Thank you so much. And it's so fun that you're not only an Instacart user, but you've actually used the ads platform. That's so cool. So thank you for sharing all of that. And I'm so glad you've had such a great experience with it. Yeah, absolutely.