Max Sinclair, CEO at Azoma, on how brands should be thinking about AI search

0:00Max Sinclair:Max, welcome. We are live on Scaling Europe. Why don't you just to kick off, give us a quick introduction to who you are and what you are building. Cool. Hi there. I'm Max, the founder of Azoma. A bit of background about me. I spent six years at Amazon working in search before I founded Azoma and we're helping some of the largest brands optimize for AI search. So we're helping the likes of Zappos, Canadian Tire, Mars, you know, the M &Ms, all of those good folks get mentioned more in AI search like Chachi Petit, Amazon Rufus and Plexi. Can you give us a bit of background? Why has AI changed search and why has this now become an opportunity?

0:47people still don't well at least on the kind of a branding customer side maybe more in the tech and vc side they do but i think people still don't appreciate how much of a transformation this is so as i said when i was working at amazon what we would used to do would be to rank keywords based on a bunch of signals so you would type in peanut butter chocolate or whatever it'd be and we would rank uh work out how to list the products based on sales velocity based on reviews etc etc like dirty signals what we have when you're using a um a large language model or a reasoning model is a totally different technology first of all these models are trained on all kind of world knowledge is as as you'll appreciate and then the um uh they have memory and kind of an understanding of the the user and their specific needs and and preferences as well so now when a prompt goes in and typically this is a longer prompt because people are kind of talking conversationally now so there's you know even more data in the prompt that feeds into the world knowledge and and also then the AI will think about the individual user as well and give a specific kind of generated response so it's an entirely different tech stack to how search used to be done on kind of or google the page rank algorithm on amazon the a9 algorithm which i kind of briefly explained and therefore requires an entirely new approach for for brands to um

2:18Max Sinclair:to surface and be recommended and how quickly or how well do you think brands are adapting to this change well i mean as i said i'm super grateful to be working with some really uh forward-thinking brands uh so there's definitely in that user adoption curve as you know there's some on the cutting edge and and we're partnering with them right now uh and and we're driving you know fantastic results uh now but i think the you know it's it's that kind of curve right so you'll see um 70 80 percent of people who will be late adopters late movers don't want to don't really want to kind of think about new technology until they absolutely have to don't want to change the way they're working and then perhaps 30 percent in those early innovators uh or early adopter kind of modes and and for those people now we're seeing great you know great rewards of of kind of moving fast and and and um optimizing for the right algorithms so right now is there a real opportunity for people to kind of adopt early and get a competitive advantage over their over their competitors yes but i think if we were talking in six months then no so i think it's you know there has been an early move advantage and we've certainly seen brands reap the reward of this but you know we've been in this three years right so I think we will soon get to the point where this becomes a new normal and I would imagine that's like six months time so like that that kind of window is is is kind of narrowing and I think you know there's lots of narrative about AI search around now and that kind of tells you that we're beyond that early we're getting beyond that early a doctor uh kind of phase yeah that's really interesting six months is a short timeline but as you said you've been building this space for a number of years how have you seen it change over the last three years how has it changed when we when we founded the business so me and my co-founder met on entrepreneur first i don't know if you you're probably familiar with this um and we met in 2022 uh pre-chat gpt and and when we as soon as we saw chat gpt you know he's a postdoc and phd in ai i am as i said i had a background in amazon at search we're like this is the future this is how people are going to discover products um that was a very controversial you know uh ef's whole belief is you need controversial beliefs and i can tell you that was a very controversial belief in 22 and we then went on tech stars in 23 of which ebay ventures were the lp and we will be sitting in the kind of ebay uh offices in in um uh kind of out in sf in the bay and we were and they were telling us yeah you're never going to have llm based search the marketplaces because it will cannibalize the ad business and the ad business is too important and this was like senior vps at ebay saying this is not going to happen so it was a pretty controversial belief in 23 and we're like no it's a better customer experience it will happen and lo and behold you know amazon of reverse Walmart or Sparky you know 60 % of us are using AI engines to search Google's now got AI overviews like we've all moved into that direction so now it's um it like we're to answer your question we're kind of we're at that point of mass adoption I think uh it's very obvious now that this is going to be the future and and kind of we're kind of thinking ahead of well what happens next um and in my mind what happens next is these agentic browsers we've um we've seen perplexity commented and if you've had a play around with that yeah um i think we're going to have these ai agents on the side of our um to start with computers but then you know on our phones which are basically running that you know hundreds of prompts in the background you know stuff that we don't have to like prompt ourselves it will just be doing stuff in the background as we talk as we search helping us to uh discover things uh kind of autonomously and i think that's you know that will be the next phase of of this ai wave and you know we're definitely thinking about that at azoma uh but obviously like very focused as well it's a current llm adoption which is just being rolled out and can you talk a little bit about the real like a real use case you know one of your clients yeah what kind of service are you providing what do they get out of it It starts with showing the customer how they are visible in AI.

6:40So we will ping the models hundreds of thousands of times on behalf of the customer. And we've got two patents actually in the business. One of the patents, we will ping Chachi Pati or Amazon or whoever it will be. We'll look at the response we get back and we build a predictive model that understands how people are talking about search. So kind of the first half is we show people how their visibility is. and then the second kind of part of it is then using those insights to generate optimized content so Canadian Tire who are let's say the Macy's of Canada you know massive retailer I think one of the largest in Canada we're doing 18 ,000 product listings with them every month with you know optimized for GEO with Zappos we're doing around 2 ,000 we have like we don't just do product listing so we have a win dixie who are a big grocer out in florida and we do we've generated for them hundreds of recipes and and as i said these recipes are based on trending topics people are talking about to chat to you so we will be picking up on the latest food trends that people are asking how do i cook xyz uh we can we'll generate a recipe that gets published on the win dixie website so then when win dixie itself will get cited which is one of the banners that uh scg owns um one of the chains that they own and then um we've increased their kind of shelf of recommendations in chat gpt and traffic from chat gpt 5x so uh yeah that's that's how we work

8:09Max Sinclair:well that's an amazing uh like testimonial or like client result like 5x is huge yeah yeah i mean we i mean i i would say we can happily 5 to 7x um recommendations and in some ways um it's you know it's not early days in terms of consumers adopting this but maybe it's you know we're in the first innings of like the GEO tool adoption from the brand so um yeah like you can you can drive pretty impressive results and what advice would you give to say a consumer facing brand who has not looked at GEO at all and maybe they're just trying to they want to get started they want to improve it maybe themselves in-house first like what would you recommend are the first steps to help them with this well my first step would be to think about who their customers are uh if their customers are gen z then 93 of gen z at least in the uk are using at least two ai tools a week so like you know if you're younger you you understand the nuances between perplexity and clawed and you're using different ones for different use cases if you are selling to an older demographic um you know maybe you're selling um you know retirement insurance or whatever like whatever it will be then maybe it's less of a pressing issue so i would i would first advise them to think about who is their target audience is that target audience using ai search and and is you know you could you can kind of tend instinctively as well as just look at the look at what data you have available and then if if their target customer is using AI search, and I would say to them, it's a massive opportunity, you know, that like likely if they're selling to, you know, under 40s, it's become a significant channel for that consumer.

9:53As we discussed, it's not yet, you know, other brands are still, you've got about six months, I'd guess, before everyone has a GEO tool. So I would encourage them to look at it and think about it and also think about Azoma as we've been doing this for three years and driven results.

10:09Max Sinclair:Yeah, of course, absolutely. And is there, you know, when you look at the models and the type of content that they kind of consume or the type of things that they're looking out for, are there specific types of content, whether it's like, you know, posting on Reddit, is it articles published on your own website, blogs? Are there other types of things that are more likely to be picked up? Yes, so definitely. And, you know, to give you the brief answer, the three things that I would encourage people to do on their own websites would be, faqs we all you know the llms love an faq if you can marry that faq with questions that customers are asking in chat gpt even better but number one faqs number two testimonials we have a couple of testimonials on on our website um and i've seen the you know i've seen these in the data and also anecdotally um being picked up in chat gpt so a really good way of um you know if you if you can put a testimonial from a real you know real customer name them that will that can get surfaced in gpt and then number three is tables these the ai engines love tables i think microsoft actually published a paper um demonstrating how like llm performance is enhanced with with a table uh when like using a tabular format so uh you know get those comparison tables up in the blogs or on your recipes or wherever it will be on the website and and that kind of you know the LLMs at least at the moment biased towards kind of consuming that kind of data as well.

11:44Max Sinclair:Interesting do you think that's going to change the way that consumer companies kind of build their websites or build their content that it's no longer almost like customer or consumer facing yeah it actually just becomes content for the models? Yeah I really I'm gonna I'm gonna give you a hot and spicy take I really just I really dislike it when people say this and the reason is in my mind these LLMs are um and it's not just you I hear people in the industry saying oh you need to design a website for the LLM right people like you have people trying to sell these these LLMs are extremely intelligent probably more intelligent than humans and they have the the mission of finding relevant information for humans and presenting it back to humans and therefore they want human optimized content they don't want llm optimized content they don't listen to llm.txt files anyone who's saying oh you need to fit that you know they're not they're not using that right they don't care they want human optimized content and if you look at how it used to work on amazon for example with the old kind of um kind of keyword uh algorithms what sellers would do would be to stack keywords so they would say i'm selling a laptop I'm going to say laptop for men laptop for women laptop for work and I'm going to like put laptop for xyz 10 times across the listing and it's just going to be stuffed with these keywords and I'm going to hit one of these narrow keywords now with the LLMs they understand context they understand the audience that would significantly punish you but actually if you describe you know what your product is the benefits of it who the audience is and you answer the relevant questions that that audience will have of it they're going to pick it up so lmlms want websites that are optimized for humans they want stuff which is easily readable they want stuff which is answering the questions humans have which is clear who the audience of you know who it's targeted for and they will do the work of giving that to the person so it's not the case that you have like something optimized for a human and something optimized say llm and these are different these are the same got it so you want you want to make sure that your content your website is fundamentally human first absolutely but perhaps in a structure yeah there's there's some like nuances which is they can't read like some java and like if you've got like flashy kind of videos they can't they can't read that and yes i i get the point which is yeah like that could be amazing for a human if you build some like beautiful 3d modeling which the lm isn't going to read but fundamentally yeah they only read text and they like tables but if we're talking about like the pure content they want human optimized content right they want it easy to read clear who it's to clear what benefits are answering the key questions um you know and the days of kind of like keyword stuffing and having like your keywords in white on a white background all this kind of stuff is is like very much over and how do you see this changing you know as you mentioned moving to like browser agents right agentic commerce do you see you know i had the uh one of our one of my first guest was philip bellum and the ceo of zilch the big big uk fintech you know they're like a payments provider but he was talking about you know they power a lot of consumer payments and he was talking about the way that agentic payments is going to or kind of like agentic checkout is going to fundamentally change the way that uh consumer websites are built because you may not need like a full checkout experience you don't need to optimize for certain things in that kind of world or basket clicks and i'm just curious to hear like do you see consumer websites and that whole interaction being built for humans like how do you see that changing as we move towards more like browser-based agentic flows i mean the again the the like the comparison i would kind of bring to this is amazon so on amazon you can have a storefront right but it's a search essentially a search bar you can have a storefront but no customer ever visits a storefront right you don't you don't like search for a product and then be like well just check out their storefront like i can tell you from working there for six years and maybe you can know as a consumer like the amount of people visiting amazon storefronts is like incredibly low and i think that will be where we get to with chat gpt i think customers will search for a product it will do a lot of research it will give you a product and then you won't go and check out the website in the same way on amazon you don't go and check out the storefront you'll be like well there's my list of three options i can read about them and i'm going to pick this one i'm going to check out in chat gpt so i would kind of agree with um you know what your previous guest was saying i i think we are going to move to that um people just you know checking out in in in llms and and kind of it will just be easier and also people are very high trust of these llms right like chat gpt has just launched a lot of stuff in health as you've probably seen and people will be uh you know spilling their deepest and darkest kind of anxieties and fears and and like secret health problems to chat gpt and it will be helping them and and therefore people are going to have an even higher trust of these models than they do now and at the moment i've seen surveys that say that people trust chanjipotty more than in-store staff and when shopping but that's going to go up with this kind of uh you know as people use it more and more for you know all sorts of um personal stuff and and then eventually people yeah be very comfortable just checking out in chanjipotty and what does this mean for consumer brands you know the brands themselves become less more less important over time and you sort of see this on amazon to an extent is that you know when you do search for a product you're not so bothered about the the brand behind the product you might look at the reviews you might look at the product specification you might look at the price do you see a world where like brand becomes less of a moat and it's actually much more important to be like a content driven content first organization where you're putting out high quality information that people want to read and understand to help you get your products into the models i think you're right i think i think some would argue that you know what you're saying about putting out high driven content is kind of building a brand and is like traditional brand building in some ways i would i mean i i would argue that brand is important as you kind of mentioned in a few um questions back about reddit and all of this stuff like you like the models do care about brand and they do want you to be cited widely on reddit and and everywhere else and wikipedia and blah blah blah so you do need a brand in order like it is important to build a brand and and put out content i think it maybe shifts the priorities of what it means to do brand building but i think it still is important got it okay so you know it's really interesting because it's always like you know you have to build a brand you have to build the content you almost have to act as if nothing is changing to an extent while also realizing that everything is changing these models and a geo is fundamentally changing the way that consumers are going to find and source your business um so i think it's like a really interesting kind of kind of world to be building in right now um like i want to hear like what are your priorities you again we talked about agentic browsers where do you see yourself focusing on as we move to mass adoption in in terms of our product roadmap you're saying yeah like where are you going to be focused either product or also geographically?

18:50I mean, we've kind of had a global view from the outset. So we've got big customers in Canada, in the US, in the UK. I have a team that speaks, on my team are people that speak Spanish, Portuguese, French. I may be missing some. So we're a team of 20 now. So we speak, we have multiple languages and we have customers across. I'd say we haven't really focused as much on kind of like mainland Europe, just because of the maybe cultural differences in terms of adopting technologies is a lot easier in the states and then maybe on a second order degree Canada in the UK so that in that will continue through that in terms of geography in terms of product I mean we see what's happening over the horizon I think if I take a long-term answer to your question I think there's two ways yet to come I think there's the agentic kind of browser and mobile experience that you know the sam altman's of the world are talking about like having this agent that helps you in every aspect of your life and we'll delegate a lot of um shopping to them so stuff like buying spare freezer parts and and like fixing a light bulb like you don't want to know you don't care you just want the cheapest and you want something that fits and you'll just you'll just give it to them to to find the products for you and then i think the next wave will be kind of ai infused products so you will have like mirrors in our house which are recommending which skincare to buy and our fridges will be saying oh if you order xyz ingredient you can cook your favorite meal uh do you want me to place that order for you so you can do it tonight and i think we're going to have like ai maybe this is like a 10-year horizon like ai infused um we're just living with like lots of ai applications everywhere as well as having like our own agent that maybe come in like two three years um and so we are you know i think we've done very well on this wave of the LLMs on on a chat GPT and and on marketplaces and yeah we're kind of thinking of the horizon as well like we want to help our customers win in these two you know two waves that are coming and probably more that I can't even you know can't even imagine yet yeah amazing and I'm conscious we're running

20:54Max Sinclair:out of time uh I want to ask you you know you've been building this space for three years yeah like one of the things I really like about you and your co-founder's background is that you have this this amazing kind of like founder market fit founder product fit you coming from Amazon your your co-founder being incredibly technical in AI. I just kind of wanted to get your thoughts, get your takes on like what advice or what kind of, what advice would you give to aspiring entrepreneurs who were like maybe earlier on than you or thinking about taking the leap? What advice would you give somebody thinking about becoming an entrepreneur or building in the AI space?

21:28Well, I remember when I was going to quit my job, I called my late grandpa and I said to him, you know, I've been working six years in Amazon. and I like I really want to do this what's your perspective you know having lived 80 years and he said to me you know go do it worst case you'll find another job in in a few years and but like at least you've tried so I would you know it was the best advice I've had and the best decision I've done to try um so number one my advice and I don't think you can do something part-time like I was you know I've all I my first business I did when I was like 16 so I've always had like like an entrepreneurial streak, but I was trying to manage at Amazon doing projects like in the evenings at the weekends.

22:12But then I also managed a large team at Amazon. It was a full on job, like, you know, eight to seven kind of job. Like it was a full on job and therefore trying to do like a few hours of an idea on the weekends or like in the evenings just wouldn't work. You need to commit yourself fully if you want to like have a big outcome or have any outcome, I would argue. and then the second piece I think is is that kind of like founder market fit I I talked to a lot of people who would who are saying oh I want to build x for y in z industry and what's my thoughts and my thoughts are there's so much you're going to be learning about your customers and the technology and the pricing and raising and like there's so many variables that you're going to have to learn and it's hard but if you're also going into an industry where you don't know anything and you're also trying to learn the industry you don't have any connections you you don't like our first customers were all of my my friends like it just you know had worked at amazon had left to you know lead the amazon or the e-commerce of major brands i had like 10 or so people that i knew over six years who'd done that and they were my first customers and they were a whatsapp away and i think if you and they can give you very honest feedback and they will do pilots with you um and if you don't have if you don't know the industry you're selling in i think my advice would be go and work in that industry for a year or two and in in some other company if you're really passionate about that industry and then start because then at least you'll you'll know it because there'll be a lot of unknowns that you'll have to deal with you know and you don't want the the the customers in the industry to be to be one of them yeah amazing that's really good advice

23:51Max Sinclair:like you know know the industry really well before you take that plunge but also I guess you mentioned like really if you're going to take a swing take the big swing yeah and go all in on it and it's interesting that you mentioned those early customers with people that you knew from your network in the the industry that that you'd already worked and i guess it's such a huge advantage three years on word of mouth is our biggest source absolutely and and you know it's like we get customers recommending us now to other people like it's it's such a and like we are the opposite of like a viral product right we are an enterprise solution of our contracts hundreds of thousands of pounds and you know it's also kind of like a secret in many ways you wouldn't want to tell your your competitors that you're using this like big enterprise software that you know but still you know word of mouth is our most most important uh one and people move around they move jobs they have friends there's conferences so i think you need to like if you don't have that you're like you're fighting against you know a cut you know uh a tsunami right like it's really hard if you don't have um that word of mouth helping you yeah it's amazing advice that you've got to build and curate that sort of like early advocacy yeah um but look max we're out of time thank you so much for joining i've loved chatting um yeah i'll be i'll be watching your journey and like get in touch if you ever want to come back on or talk about any news