Loading summary
Transcript55 lines
- [00:03]
Nick Sharma
Here's something that should bother you. Two shoppers visit your site. One looks at a $30 product, the other one looks at a $200 product they both abandon. And most brands will send them the exact same email, which is insane. Instant fixes this. It watches what each shopper does on your site and builds the perfect follow up around their behavior. The right product, the right message, the right offer and the right timing. And for the operators who care about incrementality, this is the part that matters. Instant now lets you measure performance on the same attribution window you're already using inside kl. So if your team lives by a one day click, you can see exactly what Instant is adding up on a one day click view. And believe me, when you see how much revenue your standard email setup is still leaving untouched, you're going to want Instant live. Today, over 1500 brands are live, including ThirdLove, Liquid IV and Kind Patches. And this month only, you can get 50% off your first 60 days at instant1.slash.
- [00:54]
Andrew
Sharma
- [00:56]
Nick Sharma
welcome back to another episode of Limited Supply. I'm Nick Sharma, your host and today we are bringing back an episode from the AI Summit. And the reason I wanted to do this is, as you probably know, on June 25, which is coming up here in just about a couple weeks, here we have the AI Summit. And this was a panel where we brought True Classic, Jones Road, Beauty, Unilever, and basically we're talking all about how is AI becoming practical within their businesses. So you know, whether it's talking about how Jones Road is becoming AI first, how True Classic is thinking about AI search, how Unilever is thinking about AI media buying and automation. There's, you know, this episode brings out a lot more practicality and so I wanted to put this out also to encourage you and excite you to sign up for this year's AI summit at ecomfounders.com. that's going to be June 25th, New York City. It's free for brands to attend and yeah, I hope you really enjoy this episode. This year's panels, talks, keynotes are all actually going to be very, very hands on. We're doing mostly keynotes this year and I think one or two firesides, no panels actually. And what we're going to do is basically teach everybody everything about AI. So whether it's finance, whether it's media buying, whether it's creative, whether it's, you know, website CRO, whatever it may be, whether it's setting up agents, all of that is what we're going to be going through this year. So sign up ecomfounders.com it's the AI Summit. It's June 25th in New York City at Webster hall, and I hope to see you there. It's a free ticket. If you have any problem getting a ticket, shoot me an email. It's just Nicknick co and I hope you enjoy the rest of this episode.
- [02:30]
Ben Schreiber
All right, well, thanks for being here, guys. Super excited for this panel. Let's kick things off with a little intro. Say your name, what company you're with, and a little bit about yourself. So I'll go first. I'm Ben Schreiber, CMO at Latiko Leathers. It's a leather goods and accessory brand for women and super excited for this today. Personally, I feel a bit overwhelmed by AI there's something new every day and we have the best experts here to cut through the noise. So, Alexis, take it away.
- [03:01]
Alexis Katsifanis
Hey, everyone. My name is Alexis Katsifanis. I work for Unilever Prestige, which is Unilever's premium beauty sector of the business. You may be familiar with some of our brands like Hourglass, K18, Tatcha, Dermalogica, Paula's Choice. The list goes on. I have a really fun role because I lead D2C and E marketplaces across nine brands. So there is never a dull moment.
- [03:28]
Brian Kano
Hey, guys. Brian Kano, head of marketing at True Classic Men's apparel brand. Hopefully you've heard of it. We will be launching a women's and kids line as well. So here in the near future it's going to be all apparel. But recently raised an $850 million valuation and excited to continue growing worldwide.
- [03:47]
Andrew
Hey everyone, I'm Andrew. I'm the founder and CEO of Athena. We're a technology startup based out of San Francisco. We help brands show up more on ChatGPT and AI search and drive traffic to your site.
- [04:02]
Cody
Hey guys, I'm Cody. I'm CEO of Jonesboro Beauty. We're a clean, direct consumer beauty brand. Just a direct consumer. Have nine of our own stores and excited to be here.
- [04:13]
Ben Schreiber
Awesome. Thanks, guys. So let's start high level. Every day there's a new tool, a new workflow, a new life changing automation. At this point, I think I need an AI to manage and organize my AI stack. Alexis, how do you cut through the hype and actually decide what's worth testing?
- [04:33]
Alexis Katsifanis
That's a great question. I think for us it comes back to what is going to add the most value to the customer. Prestige Beauty is at a really interesting inflection point where you have all of these mass brands trending up. We call it masstige. And so we have to convince and show and prove value more and more. And so when I think about the tech stack and implementing AI and what makes sense for our internal team members and our external customers, it really comes down to what is going to create an experience that helps them see and feel our value.
- [05:07]
Ben Schreiber
Awesome. Brian, I know you have been diving headfirst into this. You've created a bunch of automations and custom gbts. How do you cut through the noise?
- [05:17]
Brian Kano
Yeah, I think the thing that's helped me out tremendously is just not spending too much time on things. There's so many tools out there. There's so many different ways that you can apply AI. Number one is create leverage by empowering your teams to do it on their own. And then number two, if you find yourself spending more than a day on trying to figure out how to make a tool work, just move on to the next. You're going to create impact. You're going to be able to find something that's more worthwhile of your time and you can always come back to the thing that you feel. You have a high sense of conviction that actually drives impact for your business and, you know, find the time and make the space to figure it out. But yeah, for me, it's just ruthless prioritization moving extremely quickly on. If it's not working or if I can't figure it out, I'll punt it to someone else and I'll just move on to the next thing.
- [06:02]
Ben Schreiber
Yeah, it's very cool. I think it's definitely important as well to keep all that organized and prioritize what you want to do. I know myself, I have a hundred things I'm looking to get to, and every day there's something new. I'm sure a lot of you in the crowd feel quite overwhelmed in terms of what's legit, what's new, what do I need to be trying? Cody, how do you feel about cutting through the noise and organizing how Jones Road approaches AI?
- [06:32]
Cody
At least for myself personally, I don't feel like there's, like, there is a lot of noise, but it's also exciting. Like, to me, this is the most exciting time ever. And if somebody is not curious about it and like, extremely excited to be building right now and just all the opportunities we have, like, they should probably find a new profession. So there is something new every day. Like if you're on Twitter, you'll see like a New AI product launch and announce their funding round. And sometimes I'll check it out. Sometimes I'm like, hey, if this thing is really cool, I'll see a bunch of other people start talking about it. So I think that's it is like, let's wait for some social proof. Let's, let's see other people that are exploring things. What can we learn from them? And then you just have to be curious. And again, I try to like shield my team from it. I'm not going to send them every single tool. I try and say, but like, definitely playing around with things both for Jones Road and in personal, you know, life. And then when we can find a use case, then it's like, all right, how can I show this to the team? How can I encourage the team to start using this?
- [07:25]
Ben Schreiber
Right, let's dig a little bit deeper. Because you had a tweet that recently blew up. You said, has anyone else pulled to Toby and forced their team to go all in on AI and D2C? That tweet blew up. It got a lot of attention. Let's go deep. Like, what pushed you to get to that point? How's it going and how's that changed the way that you run Jones Road?
- [07:48]
Cody
Yeah, so we did. We did. I essentially followed what Toby did and a few other like tech leaders have done. And again, like, I don't think by any means we're a tech company. Like we're a physical product company. We make, we make our money selling physical products and we're actually expanding more into retail in our own stores. But in terms of just the technology to fuel our operations and back end and everything, like, it can't be ignored. Like, I fundamentally believe that like not using AI today is like the same handicap as like using direct mail instead of email. Like, it's just a basic tool that everyone has to be using. And I think in two years or five years we'll look at back on it like that. So I want to make sure everyone on my team has the, the tools and the knowledge to be able to do so. And then obviously, I also think, you know, as we grow, there's a lot of disadvantages of scaling a D2C brand and operating leverage is really like the one thing that you have. And so I want to stretch my current team to not necessarily cut headcount now, but as we grow revenue over time, how do we do so as efficiently as possible? And I really believe this is the opportunity to do it. So I wrote an email, was a super long Email. I'm sure half the team didn't read it, but just like we're going AI first, this is why this is how we're thinking about it. I wanted to also talk about the elephant in the room, which is job loss and replacing some of them because it's definitely on people's minds but show how we're thinking about it, which is we're not trying to cut current jobs, we are trying to help people get more productive in their job. Better quality of output and then as we grow in the future be able to do so while adding headcount less. So shared our plan, how we're thinking about it and then also did a little bit of a training. So I did like an hour long training. Again, I'm no expert, but just sharing some cool stuff because the thing I realized is like, you know, we are all obviously people are at a conference. Like founders are constantly thinking about this. Like I'm building stuff on a Saturday for fun. Like most people aren't doing that. And so yes, everyone knows ChatGPT exists and I would say half the people were using it daily, half the people weren't. But like not everyone knows about deep research, not everyone knows about Vibe coding and all this stuff. So I just wanted to like kind of blow people's minds a little bit. Everybody has probably had that experience and everyone here I'm sure where you just get red pilled and you see something and you're like holy shit, this is like changing the game. So I just wanted to show that. So showed a bunch of examples and then we're bringing in two to three like expert speakers to train the team and provide kind of like best practices. So bring in somebody like Alex Cooper is going to do a training for us on just general AI workflows and building agents and stuff like that. And then we also did a bounty. So we're doing a $10,000 bounty end of the year. Just was like, hey, whoever builds the coolest shit with AI that has the biggest impact on the business. Um, and that was really cool. Cause the next day like our procurement manager sent over like two calculators that she built and Vibe coded that like was a ton of manual work ahead of it. So I'm really excited to see what people build and I'm sure that $10,000 is going to be like a really good investment because we're going to get unlock way more value out of it.
- [10:39]
Ben Schreiber
That's super cool. The bounty to whoever creates the best automation or agent is super cool. I'm sure. People are coming up with some awesome things to help Jones Road. I just want to double click on one thing which is, you know, a lot of the employees will obviously be concerned about job loss, automating themselves out of a role. How do you ensure like that they know this is to make the organization more efficient and better as opposed to, you know, automating themselves out of a role, whether that's creative direction, copywriting, whatever that may be.
- [11:14]
Cody
Yeah, there's probably what, what I say and then, you know, I can only hope that they take it at face value and I think I have to make sure that our, our actions as a company align with our words. But yeah, I, I fundamentally am not trying to. We, we would like to be more efficient over time as we grow. We think this can help again make, help people get more productive in our role, in our role, you know, current role. So we have one copywriter right now for the whole organization and instead of hiring a second, I would rather pay our current copywriter 20% more if she can then be 10 times more effective and efficient with AI. So I would actually prefer to keep our current team, pay them more because their value that they can provide to the company is now so much greater. I think there's, that there's obviously automating busy work and so there's a lot of manual things and I think we've been a little bit stuck behind with technology, but there's a lot of manual things that can easily be automated to allow team members to actually focus on higher level strategy things like that. And I did give a little bit of a wake up call which is like everyone is probably going to have another job after Jones Road and like the way that work is changing is totally going to change and these lower level tasks are potentially going to go away or get automated. And I highly encourage everyone to get really good and like just know where, skate to where the puck is going and know that like the way we work has to change and instead of managing people, we're going to manage these AI agents and we, in order to be valuable in this new world, like it's much more of a strategy role. Yes, you have to understand AI tools and use them, but like you can't. There's always going to be people that are needed for the strategy and kind of managing the overall direction and so just encouraging people to kind of work on those skill sets as well.
- [12:55]
Brian Kano
If I, if I could just add, I think that last point's the most critical because it's less about the adoption within your organization, but it's just how the industry and how society as a whole is moving. So get paid to learn, get paid to do really cool stuff, make the organization better, but then take those skills in to your point earlier, apply those at the next job and your next career versus being the person that is kind of like waits and then falls behind in the industry.
- [13:21]
Cody
Yeah, agree with that. Yeah, I also wanted to make it clear, like we're also investing in people and like, kind of like Google, I don't know if they still do it, has that like 20% time where they can play around with stuff like giving people budgets for softwares and encouraging them to actually spend time on the clock where they're playing around with stuff. Hopefully it's Jones Road related, but they're also working on those skill sets like Brian said, which hopefully helps them as well.
- [13:41]
Ben Schreiber
So let's switch gears for a second and talk about AI search, which I know is on everyone's mind. Andrew, you're on the software side with Athena tackling AI search. Why don't you give us just like a 30 second rundown as to how you're tackling AI search and what you're seeing brands doing that are having success.
- [14:03]
Andrew
Yeah, happy to. And Cody used to. Yeah, Google. And that 20% time is still very much a thing.
- [14:11]
Cody
So
- [14:13]
Andrew
what we do, Athena, we work with a lot of brands across different verticals. And in E Comm specifically, we see that we're at this cusp of bridging the gap between research and shopping on ChatGPT. Just to get a quick pulse, check how many people here are using ChatGPT on a daily basis? Okay, that's a lot of hands. How many of us are aware of ChatGPT shopping coming out? Fantastic. How many is aware of Perplexity Checkout using Venmo coming out this summer? Okay, so some background perplexity is adding Venmo checkout onto their search console where anyone can check out on Venmo. And the implications this has on brands is that these AI search engines are going to be in some cases bypassing your website. And if you spent the last 5, 10 years optimizing your website for humans, you just start thinking about how you optimize your website for AI. And that's how we're thinking about this problem is that humans, we understand content differently than AI does. Right. AI has different biases towards different structures of content, different sources of content. Backstage right now we're talking about Reddit and how Reddit is one source where we're Seeing brands seeing the most success is where they're able to be very targeted in how they approach AI search. I'll give you an example for Reddit. A Reddit strategy could be not just targeting, you know, blanket firing, shotgun, approaching Reddit as a whole, but using a tool like Athena to be very pinpointed on which subreddits you should be caring about, who are the exact users on Reddit, the exact comments that are influencing AI for your product, for your vertical.
- [16:10]
Ben Schreiber
Awesome.
- [16:11]
Andrew
Thank you.
- [16:13]
Ben Schreiber
You know, Alatico, we started diving in head first to AI search about a month ago. We're seeing about a 10% increase in brand mentions across our top, middle and bottom of funnel. So we're really trying to be first movers there. The Chat GPT Shopify integration, where you can check out within ChatGPT is in beta and so that's coming soon. The Perplexity Venmo integration. Super excited about that. So I think for any brand this should be a first priority. Brian, do you want to talk a little bit about how you guys are approaching AI search?
- [16:51]
Brian Kano
Yeah, definitely. I mean number one is we're starting to actually track it now. I think over the past in Q1 I started tracking and I see a couple sessions and you'd see like seven visits, 14 visits. But I want to say around April, April 12, we just started to see this hockey stick growth in terms of sessions, new customer purchases and then we, the platform we use, we use triple. Well, so we can distinguish between first CL and last click or any, any journey. And shockingly, a lot of those conversions are actually first click. So this is what's now forced my attention to say, okay, I need to get in on this and sort of amplify this and 10 exit, because there's momentum here. So one of the things that I'm looking at is one, understanding how and what search engines or sorry, AI is using as a reference or a citation to make a recommendation to a user on product. The one thing that I'm seeing that is frequently cited is Reddit. And so right now I have an initiative with my team that every week we're going to give $500 of store credit to Reddit users at random. And I have a specific criteria, but these are going to be people that are active, they seem to have a lot of trust in their local community and they have expressed some sort of desire to a problem that we solve or we just want to make their day. This is someone that's very wholesome and it's aligned with our brand. Value. So let's just make their day. So we're giving away the store credit, and this is an early initiative, but the goal is to take these Reddit users. They're already active in the communities that have what I would consider domain authority in the space to get them to talk about the brand. And then over time, we end up having a widespread volume of citations that AI can reference and hopefully pull into and responses around. What's the softest tee? What's the best T shirt? Is it durable? What's the best price for value? And those are ideally quotes that are in the post itself. Now, Andrew and I were talking about whether or not it's a viable strategy, and he says, yeah, it's definitely a viable. But one advice that he gave me was being very intentional, because it's one of the problems I face is that $500 does not go a long way when you're talking about a Reddit community.
- [19:08]
Andrew
It's.
- [19:08]
Brian Kano
It's a lot, it's massive. There's so many different communities. So Andrew's feedback to me was get more intentional about which communities specifically have the highest engagement, the highest domain authority, and which ones are going to be the ones that actually rank you up higher so you don't waste your time and energy in some of the communities that aren't as high of a consideration for AI.
- [19:30]
Andrew
Another thing I'll add is that the user base who are coming from ChatGPT, I'll send a spat, which is 90% of people who earn $125,000 or more a year are using ChatGPT. And what we're finding across the board with companies we're working with, is that the types of users are getting from ChatGPT, they're converting at a higher rate, they have lower bounce rates, they're higher engagement, and that's because they've gone through this whole buying journey on AI search before even landing on your website. So the types one visitor from traditional Google is worth less than a user that's coming from ChatGPT because these users are already going through this journey that you might not be aware of. You might not have any way to track it.
- [20:14]
Ben Schreiber
Awesome. So when brands are strategizing for AI search, Andrew, what can they look for in terms of key strategies, pitfalls to avoid? And what are some metrics that matter when you're looking at geo AI search compared to traditional SEO?
- [20:33]
Andrew
Yeah. So I'll give you a tip that's generally applicable to everyone here, which is when you go back after the summit and go Back to your teams. Make sure you look at your robots txt files. It's a very simple fix. A lot of companies I see where they shoot themselves in the foot is where they're blocking the bots from crawling the website. That's a very simple fix, very simple mistake you could be making, but is causing a website to not even be picked up by AI search. So that's a very general tip I can give in terms of other advice I'd have for getting onto AI search. I think it really is about being very targeted. As Ryan is saying about every company, you have a digital footprint on the Internet. And the difference between AI search and humans is that on traditional Google, I think 70, 80% of humans click on the first three links. We know that from the old SEO days. And if you got to the first three links, ideally number one in the keywords that you care about, job's done. With AI, what we're finding is that across all the companies we're working with, they're seeing hundreds of different websites that are influencing the AI responses and the distribution of how influential they are is more so. Even so, there's not this like skew towards the first three websites, but rather it's the first page of Google, the second page of Google. All of these websites are influencing the AI responses. And what that means is that you should be looking into not just how you're ranking but how you're showing up on these off page third party websites which are also talking about your brand and AI is trusting them as well as these kind of thought leaders. Right. Because humans, we have our own thought leaders, AI has a different set of thought leaders and uncovering who the thought leaders are, thinking about how you can influence these third party websites, that's another strategy.
- [22:36]
Ben Schreiber
Yeah, it's one thing that stood out, I think you mentioned this to me, was like with Geo, it's not whether you're mentioned, it's also how you're mentioned. So it's a different way to think about it and to strategize when you're thinking about AI search. I also read a crazy stat which is that 50% of traditional search traffic will be replaced with generative AI by 2028. I think maybe that is too far out. I think it's coming a little sooner, if not sooner. Yeah, yeah, cool.
- [23:05]
Nick Sharma
Time for the retention tip of the week brought to you by Instant quick gut check for every founder listening. Do you know your real repeat purchase rate? Not the Shopify dashboard number that includes subscription renewals and auto charges. I mean your real organic repeat rate customers who actively come back and place another order on their own. Most founders I talk to overstate their repeat rate by 15 to 25 percentage points because they're still bundling subscriptions. That number matters, but it masks whether your brand actually has pull. The organic repeat rate tells you if customers love you enough to come back without being on autopilot. Here's how to find it. Pull your total repeat purchase rate, then subtract subscription renewals. That delta is your real retention. If it's below 20% you've got product market fit issue or a post purchase experience issue. If it's 20 to 35% you're in a healthy range. Above 35% you've got something special and you should be investing way more into retention infrastructure. Once you know that number, you can start building smarter flows to grow it. Instant helps brands do exactly that by combining AI generation to identify and re engage shoppers who might otherwise never come back. Check it out at instant1/sharma for more.
- [24:12]
Ben Schreiber
So switching gears to more so like how this AI impacts media buying analytics. You know Brian, specifically a true classic. You probably spent more on creative testing than many brands spend in the entire year period. How are you approaching implementing AI in your media buying and and add creative strategies?
- [24:33]
Brian Kano
Yeah, I I think of AI use and applications in three buckets. My brain just works in systems and so I love to like bucket and label things but number one is operation. So the thing that we're trying to solve for right now is how do we operationalize media buying. And that can be in the sense of uploading the ads, changing the bids, changing the budgets, identifying fatigue and calling that out and sending updates in Slack. The other is an insights which is actually predicting and giving some sort of root cost analysis into why something is working or not working. That's another thing that we're trying to work on right now we are probably like 20% there, right? We're using a lot of data and it's very manual. And the vision that I have is something that's proactive rather than me having to prompt and upload a spreadsheet and say analyze this and then me kind of coaching and being the guardrail and I want there to be kind of a 24. 7 on at all times AI employee that is flagging hey CPM deviations cost per click. This ad has been identified as having fatigue because over the past three days the click through rates decreasing 20% day over day. All of that proactive communication where it doesn't require any prompting. That is the vision that I'm working to bring to life, but we're really behind on that one. And then the last one is Generative, which is just like making the ads. So where I'm spending a lot of time right now is on sort of a combination of Insights and Generative. Right now we have operations, I have teams working on operations. So as an example, we have an AI agent that views all of the creative the ads that are uploaded. It checks for URL links, it checks for tracking parameters, it actually views the ad and then it goes to the website and takes a screenshot and it sees if the offer and the product is matching. And if it doesn't match, it'll send a slack message. And then the media buyers have 24 hours to address that issue. Either say, yes, I resolved that, my bad, I made a mistake, or this is actually not an error. So we're, we're operationalizing that. We're getting to a place where media buying is going to be automated here hopefully in the next six to 12 months. But where I'm spending a lot of my focus is Insights and Generative, mostly on the creative strategy. One thing that I highly recommend you all to do, download your sales data. Sorry, Download your Shopify sales data by zip code. Then download your spend impressions, reach and performance on meta by state and then break it down by age and gender. Then download all your post purchase survey data. You're going to take all this information, you're going to go and create a GPT or you can just do it in chatgpt yourself, upload it all and then tell ask AI to identify 5 Personas or 10 Personas for you and bucket them into opportunities and sort of like a core. The core is where you have a high volume of saturation, you have a high volume of exposure and performance is good. And then the opportunities are the one where like volume is really low and opportunity and the conversion rate's really high. So an example might be like Philadelphia, maybe you don't spend a lot in Philadelphia, but the conversion rate's higher than what you spend in California. And due to the zip codes, you can make some estimations around who this person is, what they're buying, what their age and gender is, and even like which channels are driving visits to the website. So with those Personas, now you can take all the information and use a tool like icon or veo or chad itself to just go and start making ads that speak to those Personas. So right now it's very manual Right. I'm downloading the data, I'm creating the Personas, I'm having to interact. But it is helping me now analyze large amounts of data to then guide creative strategies so we can identify new headlines, new opportunities for us. For example, we identified that blue collar workers, delivery guys that, you know, work for UPS or Amazon, what they, they go through T shirts quite frequently. It's like the base part of their uniform. And so we started to create ads that speak to this Persona and it works. Right. So that feedback loop, I think right now it requires a lot of human involvement, but it's very actionable. You guys can go and do this later today. But the vision is how do we start to automate all that and plug it together so it just feeds itself?
- [28:55]
Ben Schreiber
Yeah, I mean, I think the whole process, right. Is getting there and it's evolving super rapidly. We just started using the triple whale AI agents which allows us to, you know, use that automation in many aspects of our media. Buying, creative testing, creating prompts for our creators based off what our winning ads are, what our winning hooks are and all those metrics. How are you guys approaching this at Unilever?
- [29:25]
Alexis Katsifanis
Yeah, we have two main work streams happening at Unilever. The first is creating value add experiences, which is really table stakes. So for example, Paula's Choice is our largest D2C brand in the portfolio. They started in the 90s, so arguably maybe the first D2C beauty brand. And they have all sorts of homegrown tools which is, is really fun. And we swapped out their home built quiz online with RAVIV's AI AR powered skin diagnostics tool and we saw a 15% lift in conversion rate, 30% higher AOV and it drove around 3 million in the first couple months. And what was great about that, of course the technology. But for customers it felt more like a give get than the original solution because we were educating them about their skin or if I look at a K18 hair, you know, they're understanding why. And the humidity, it's starting to go like this. So it kept them coming back and then we could take their results via the selfie upload and use that post purchase and a loyalty program to help them better understand tracking changes over time. Wow, this regimen really is working for me. I'm going to tell a friend. We similarly at Tatcha are updating some of our chatbots. So we just onboarded Alhena and we're seeing like a 50% lift in AOV from that, which is amazing. And that I think something that hasn't been discussed yet on this stage is the challenge with brand and AI because we have really specific brands and prestige voice and tone. What we're willing to say, what we're not willing to say, even in a chatbot, in a quiz or copy on site, for sure. But when it comes to AI search, which is our other work stream, of course we're doing the tracking and understanding from our web blogs. You know, what crawlers are coming to site, what it's, what it's driving for us and our share of model across the different LLMs. But then what to do with that information can become difficult for us when we know we need to build copy for AI crawlers and not humans, when we have a very specific voice and tone. And so I think about an Hourglass, which is a luxury cosmetics brand. How are they developing copy? How are they going out to third parties? And what are they willing to say or have folks say about them? That influences the LLM that still feels on brand is a challenge that we're currently looking to tackle.
- [32:02]
Ben Schreiber
Super interesting, the personalization aspect of AI, especially for beauty brands, specifically, what led you guys to start doing that? And for any brands in the audience, how would you recommend them to take the first few steps in regarding to personalization with AI, you know, regardless of which industry they're in?
- [32:21]
Alexis Katsifanis
Yeah. Again, I keep coming back to the customer and for us, what we needed the customer to understand was what's best for their skin. In a world where everyone is a TikTok influencer and you're hearing about beef tallow and then the same time you're hearing that sunscreen is the devil and we really wanted to cut through the noise. So we look at different AI solutions and personalization that can help us cut through the noise. I think if you have your one KPI, that's actually not a brand metric, but a customer metric. That's at least the way that I'm looking at it today. And that's how I determine what works and what, what won't work for us, at least in the interim. And then of course, it's the forever test and learn and optimize. I think in my role, what's unique and fun and hard is that the brands are at different stages. So A Dermalogica, for example, has been using AI since like 2018. They had built a custom model that was training their skincare therapists on how to just better educate customers in the spa to understand why they should purchase these products after their visit. And then today they've built their custom tracker and they're delivering like 80 plus prompts a day to understand where they have the right to win and what are some specific areas they need to double down on like retinol or microbeetling. So I think whatever industry or area you're focused on trying to understand where you have the right to win and how that will help your customers keep coming back and give them a reason to keep talking about you.
- [33:55]
Ben Schreiber
Awesome. I want to end on one last question for everyone here on the panel. If we're sitting here one year from today, what is one AI tool or capability you think will be absolutely essential for DTC brands? Cody, let's start with you.
- [34:11]
Cody
And that's a hard one. There's going to be a lot, I don't know how many, like universal ones, I think to not answer the question and well, the end, like there will be so many different vertical ones. Like I think AI will, if it's not already, will impact every area of the business. So influencer, there will be amazing solutions that'll automate most of influencer marketing and a lot of sourcing and things like that. There'll be AI stuff in finance ops as well. So everything. But I think a lot of those will be much more vert vertically integrated sasses versus just one overall. But if, if you got to pick one, just pick a model, pick Gemini ChatGPT. Claude, just get really, really good at using that because that'll just be like, like we use email, like, we use, you know, Slack right now.
- [34:54]
Ben Schreiber
I mean, I think that's a great answer. It's, I think you have to be doing a lot of things. You have to be arming your team, you have to be staying on top of trends, seeing what's working. And so Brian, what would you say
- [35:08]
Brian Kano
12 months from now? I think that AI will be more proactive rather than reactive. If you think about all of our use, all of our examples and how we interact with AI right now, it's driven on a human needing to ask and prompt and then get something back. I think in 12 months that's going to completely change. It's going to be AI telling us, hey, I identified these risks, these threats. These are your priorities for this week. Go and do it right. And it's going to be telling us what to do. That's 12 months, I think 10 years from now. I don't remember who said this or where I remember reading this, but it resonates so well, especially hearing Andrew speak. I think the Internet's going to be one Giant newsfeed. I see this world where all of the personalization and how people engage online is just going to be one steady flow. And if you think about what made Facebook so successful and it was finding the right product for the right person at the right time, I think AI is going to scale that out beyond just the social walled gardens. And the entire Internet is going to be this layer of product recommendation for the right people at the right time in their life. And so 10 years from now, the thing that I think is exciting and maybe I'm going to find ways to invest even more and is this, this concept of geo and sort of preparing for a world where AI is driving all of the sort of online behavior for users.
- [36:37]
Ben Schreiber
Awesome. Thank you. What do you think, Alexis?
- [36:40]
Alexis Katsifanis
I think this is a really difficult question. With the rate at I wasn't going to take it easy at the moment. I think I would say I would love one AI tool that could train my teams on the most important things to know and understand about AI, because I think it is, to Cody's point earlier, really exciting. But for the teammates, it's. It's like too much of a good thing at once. And I want them to understand where to focus. And if we could just spend one, one day, one thing, training everyone and then move forward from there, I would love that. And maybe it exists and I just don't know. So I would love to know if it does.
- [37:16]
Ben Schreiber
Cool. Then lastly, Andrew, I guess specifically with AI search, how can these brands be prepared for what AI search is going to look like 12 months from now?
- [37:26]
Andrew
Yeah, well, I'm a little bit biased, of course, but some background on myself. I used to be a product manager working on Google Search and I was working on the information acquisition team. And I can't say too much, but I can say that, you know, if you like to check out the latest demos of Google Search's newest features, it's going to be a lot more agentic, meaning people were already using, you know, most of this room is using ChatGPT already in 12 months, I think a lot sooner even, we're going to have each person, each consumer is going to have their own personal shopping agent. And the shopping agent will know their personality, their preferences, what they like to buy, what the previous brands they bought from, and then it'll be going out there onto the Internet, making autonomous purchases on their behalf. And in that landscape, that's what we're building Athena for. That's where I see the, the Internet and commerce heading is that when humans are offloading the purchasing decisions more and more to AI and their own agents. What does that mean for brands? That's where Athena comes in. Helps your brand monitor, get insights on where your perception gaps are, take action and operationalize the action.
- [38:40]
Ben Schreiber
Awesome.
- [38:41]
Andrew
Thank you.
- [38:42]
Ben Schreiber
I think that wraps it. Hopefully you guys got some valuable takeaways from this panel. Alexis, Brian, Andrew, Cody, thank you so much for your time today.
- [38:51]
Alexis Katsifanis
Thanks Ben.
- [38:54]
Nick Sharma
Thanks for listening. We'll be back. Next time to cut through the noise on cpg, retail and e commerce. If you enjoyed this episode, why not share it with a friend? And be sure to subscribe wherever you listen so you don't miss the next one.
- [39:12]
Cody
Sam.