Sylvain Utard – General Manager AI at Algolia (https://www.linkedin.com/in/sylvainutard/). Algolia Acquires MorphL, Launches AI-Powered Predictive Experiences and Personalization: https://www.globenewswire.com/news-release/2021/01/26/2164201/0/en/Algolia-Acquires-MorphL-Launches-AI-Powered-Predictive-Experiences-and-Personalization.html
#19 Sylvain Utard – General Manager AI at Algolia on acquiring MorphL and the future of Algolia's AI products – Get Your AI On!
Ciprian Borodescu: I’m here with Sylvain Utard, General Manager of Algolia, overseeing Algolia’s AI offering to bring intelligent search to dozens of thousands of customers. I’m super excited, and it’s an honor to have you on this podcast. Thank you so much for being here, Sylvain.
Sylvain Utard: Thanks for having me, Chip.
Ciprian Borodescu: By now, the cat is out of the bag. And so, I think we can share the news with everybody. Algolia acquired Morphl at the beginning of this year to help power Algolia’s new AI offerings. And this is exciting in more than one way, but before we dive deeper into the details, a few words on Algolia. Algolia is the search as a service platform that enables companies of all sizes to deliver fast and relevant digital experiences that drive real results. With Algolia, consumers are able to easily find and discover what they want across the web, mobile, and voice. Algolia allows developers and business teams to build and optimize delightful search and discovery experiences that increase online engagement, conversion rates, and revenue. Now Sylvain, let’s start with the beginning. You have been with Algolia since its inception and if I’m not mistaken, you are employee number one, right?
Sylvain Utard: That’s correct. I actually joined them while they were about to raise their first seed round. That was definitely early in the days of the Algolia journey. That was in September of 2013.
Ciprian Borodescu: Okay, so do you remember the interview or the conversation you had with the founders more than seven years ago? How did that come to be? And how did you make the leap?
Sylvain Utard: I definitely remember this because, of course, it wasn’t really an interview. I’ve been working with Julien and Nicolas – the two co-founders – for a few years before Algolia. And so, when Julien quit this company that we were working together for and started his own company, he told me about what he could do, what it might do. And along the way, he finally told me that Algolia is gonna build advanced libraries with high algorithmic value for mobile devices. I was like, “Okay, mobile devices, you know, Julien, this is not really what I’m the best at. Really, I wish you the best of luck, but I’m not sure I would be the right person to transition at you.”
Ciprian Borodescu: Oh, really? You did say that? Oh, wow, okay.
Sylvain Utard: Sylvain Utard: I definitely said that.
Ciprian Borodescu: Was it a strategy? Or did you play hard to get or…?
Sylvain Utard: Not at all. Actually, at that time, you know, I was really passionate about the web. I was really passionate about building high-performance engines, but really going to the mobile world, this is not where I’m coming from. And really, I had zero experience on mobile. And about six months later, he came back and told me, “Sylvain, you know, what? We’re actually pivoting the company. The mobile business is tough because, in order for us to convince a mobile app to use our technology, it’s really about connecting with all these mobile developers, most of the time these are agencies. It’s really kind of a long journey in order to get to a decent amount of customers. Therefore, we are pivoting to a SaaS model where, with the same technology that we built – and at that time, their first high algorithmic value library was not your search engine – we are going to take this search engine that we have been designing for the mobile – that at that time, by the way, was running super fast on the cheapest phone that you could get; and back in 2013, trust me, you couldn’t really use much CPU or RAM or OBS. But they did a really good search engine for mobile devices. And the new pitch of this type of company was, let’s take this technology, host it on the web, and expose an API to build a SaaS model.
Ciprian Borodescu: Interesting.
Sylvain Utard: Fun fact, I actually resigned from the company I was working for two weeks before Julien reached out and so, my first answer was like, you know, “Fun fact, I’m actually free, and I was about to start looking for a new job, so this is actually perfect timing.”
Ciprian Borodescu: Exactly. Interesting. Walk us through your journey at Algolia during these last seven years, and maybe share a few happy or challenging moments and how you managed to overcome them.
Sylvain Utard: You know, this is something that I keep hearing in all startups so I think we’re really not different from the others. In the very early days, you are only a few people, usually, all of us have an entrepreneurial mindset, all of us are willing to do way more than what we’re really trained to do, meaning that sure, people listening to this podcast will know that even if you’re an engineer, you might do marketing, you might do sales, you might do product management, and so on. So, at the very beginning of Algolia, I was and we were all individual contributors of this product that we are still selling today, the Algolia Search API. Of course, we were lucky enough to join Y Combinator, the winter 2014 batch. And then, lucky to have a few great names that have been enabling us to start this hyper-growth that we have been experiencing. And when you are in such a hyper-growth mode, the idea is that what you do day to day is really changing every day. This is something that I’ve been surprising myself to say oftentimes, but even today, it’s quite accurate. The team is growing, the company is growing, the product is evolving, and all of that makes it very hard to keep doing what you were doing six months ago because you need to scale with this hyper-growth and therefore, as a contributor to this success, you need to adapt and to literally change the way we’re working.
Ciprian Borodescu: Would you say that YC was an important milestone in the growth of the company? I would imagine so, but was it like an explosive moment?
Sylvain Utard: I think we can definitely consider this as an explosive moment, because, you know, we started this company in France and we knew that our core markets were all these websites that have a tech team, all these apps that have a tech team and that we’ll be able to integrate our Search API. What we do, at the core, is really a tool for developers. And so, there is one single place where developers are in huge quantity, which is Silicon Valley. And so, we really thought that was the best for our success to go to California and we were super happy to be accepted within this YC batch from which not only we have been learning a ton at YC – there’s one thing that I think I will remember: they really challenge you to get 10% growth per week. And you choose the metric. Depending on what you do, you might want to use ARR, but some other companies measure their users or the searches. For us, it was the ARR and 10% a week for three months, this has been, I think, what really put us in this mindset of hyper-growth. The reality is that once we came out of YC, of course, this YC Stamp that we had, we came to use it a ton to convince our first developer customers in the US. And once we had those big names, then it was way easier to think about them rather than thinking about the accelerator.
Ciprian Borodescu: And I can testify to that. I mean, we were through Techstars, and definitely, when you say that name – YC, Techstars, 500 Startups – you know, the top accelerators in the world, they really open up a lot of doors. And the network is really, really important.
Sylvain Utard: You were asking about, you know, happy or challenging moments. I think these three months in a house that we rented in Menlo Park, six of us – the two co-founders, our initial VP of business development, and myself, plus two interns – this was an amazing, but also challenging moment. Such a cliche, let’s be clear; we were in this house, working all day long, but…
Ciprian Borodescu: Yeah, in a garage or something like that. Yeah.
Sylvain Utard: Yeah. Kind of. Working in the kitchen, and, you know, trying to make it work with six people in a single house. This was definitely a happy moment and a challenging moment. I think what I keep also in mind about these past seven years, the offices that we opened around the world, we now have an office in San Francisco, Atlanta, New York, London, Paris, Tokyo, and recently, Bucharest. So, all these milestones are something that I really liked. Of course, the current pandemic situation is making it a little bit hard, and most probably the offices will never be the same as before. But still…
Ciprian Borodescu: Excellent. Do you remember any other startup names from your YC batch?
Sylvain Utard: Definitely. Not sure you heard about them, but Flexport – they are doing very well. One batch after us there was Firebase. They are doing databases themselves.
Ciprian Borodescu: Yep.
Sylvain Utard: They were later acquired by Google, but it was very similar. And then we had… It’s a long time now so I tend to forget all the names, but we stayed in touch with a few founders. We actually had the privilege of having most of our YC peers Algolia customers at the end of the day.
Ciprian Borodescu: Oh, nice.
Sylvain Utard: They were so happy with the Search API that we were providing them at that time, that some of them have also joined our safe round after YC. So, some of those founders are also business angel investors of Algolia.
Ciprian Borodescu: Oh, that’s interesting. Very interesting. I didn’t know that.
Sylvain Utard: It was a very nice challenge for us. We were like, okay, if we want to build the best product of a tech company, we should aim to have 100% of all the YC companies that are in our batch. We have direct access to them, they have our trust more than anyone. If we’re not able to convince them, we are doing it wrong. And so, I think this extra challenge that we put ourselves in was very successful in the end, and kind of helped us design the right product.
Ciprian Borodescu: I’m sure that our audience would like to know more about something that is public now. Algolia acquired Morphl and there’s a lot of excitement around the common vision. But before that, maybe you can walk us through some of the AI initiatives you previously had at Algolia, and what were your key takeaways?
Sylvain Utard: Maybe one thing to keep in mind is, at Algolia – our website is showing this – we are building a tool for developers and we are developers also. And so, there is one thing that we have always been doing and it’s aligned with our core values: we really say what we do and do what we say. And so, for a long time, there wasn’t much AI on algolia.com. Yes, we were doing some machine learning, we were doing a ton of statistics, and you know, tie-breaking systems and decisions like this. But really, for us, as engineers, it wasn’t real AI. And so, for a long time, we haven’t been speaking with AI. And, at some point, we believed that this was maybe not the best approach. We should be more proud of what we are building and, you know, behind the AI world, there is a lot that can fit. And really, what we started with was a way for our customers to continue getting more and more value. And we have been doing three main features, I would say, so far, in this AI scope. The first one has been this personalized search experience where – and you know that’s why you are going to Google, it’s kind of mainstream for everyone now, but if I’m searching something on my Google and you’re searching the exact same thing on yours, we might have different results. This is because of where we are, this is because of what we have been searching before, what we have been clicking before maybe. All of that is something that is priceless for our customers. A lot of them are in the e-commerce space or the media space, and for sure, being able to personalize the results to provide every end-user with what they really want, has been really appealing to our customers. So, therefore, we built this personalization feature that you can turn on; it’s ingesting a few events, that of course, you need to send over to Algolia, and then we’re able to provide you with personalized search for you. There’s a tool that actually is going to be released later this week and it’s our re-ranking feature, which, again, leveraging all the events that we get from the end-users will take some decisions to change their overall texture and business ranking that is configured within your index to promote or demote some objects based on the behavior of the user. This is really about, you know, if everyone is clicking the sub-result, it should most probably be the first one. So this feature is definitely aligned with, the more you use it, the better it gets.
Ciprian Borodescu: Exactly.
Sylvain Utard: The last one that we have been developing on AI is a way for customers to get sign-ins in order to increase the recall of their subjects. This is about helping our customers figure out how their users, their actual end-users would search that catalog, for instance, on e-comm. You might be speaking about, you know, a pair of sneakers – maybe your catalog is written ‘shoes’, or whatever else. And this is about providing them, again, learning from all the behaviors of the users, what could be the right sign-ins to add to your Algolia search.
Ciprian Borodescu: And just to put things into perspective, let’s maybe put some numbers behind Algolia. I know that Algolia has around 10,000 customers, and this is public information. I think we’re about 100 billion queries a month, is that accurate?
Sylvain Utard: That’s actually a little bit more than 100 billion. We are currently at, if I’m not mistaken, 150 billion searches every month.
Ciprian Borodescu: My God!
Sylvain Utard: This is huge when you look at how many users are behind the searches. We have been computing that about 100 million IP addresses are using our API. And therefore, what you can tell from these numbers is that assuming there are maybe 3.5 or 3 billion users on the internet every month, literally Algolia is touching in some way one out of six, or one out of seven users on the internet. It’s absolutely massive, indeed, to know that we are the source of all of that. But we also have a ton of free customers that we are not picking in this 10,000 because again, we are a developer tool and we strongly believe that it’s very important for any developer to be able to test and play with Algolia without paying. So, in addition to these 10,000 customers, we probably have – I don’t have the exact number – between 30 and 50k free users that are able to implement Algolia.
Ciprian Borodescu: Okay. So I think we prolonged it enough. Let’s address the elephant in the room. Algolia acquiring Morphl. And I can share my side of the story because it’s pretty straightforward, and then you can share yours. So it was October 21st last year, 2020, the year of the pandemic, when I noticed both you and Julien – Julien being the CTO at Algolia – signed up at Morphl. And of course, I knew about Algolia. And the first question that popped into my head was, “What is the CTO and the VP of Engineering at Algolia – basically, the executives of a big company – doing in our platform? And what are they up to?” Then the next thing I knew I was sending you, guys, an email to give you a demo. And honestly, I did not expect you to reply, but you did. And you replied, like, in two minutes. And so, our conversation started. And although we weren’t initially in the right mindset to be acquired because we were in full fundraising mode, we had very, very similar visions, and it was simply the right thing to do. And we’ll get back to the grand vision behind this acquisition, and the products we’ll be launching soon at Algolia, but I would like to hear your side of the story. And just for our audience to know, in fact, this is the first time I’m hearing your side of the story. I’m actually pretty curious about it.
Sylvain Utard: Yeah, you know, at some point, by realizing these three AI features that I was describing, we were really only scratching the surface. And so, what we wanted to do is figure out ways for us to go fast. And at our stage, a good way to accelerate and a good way to, you know, not only build a better product, but maybe also increase your total addressable market, and also unlock maybe tenfold, is an acquisition. And therefore, of course, we have been looking at various companies and there were a few things that we were really looking for in the companies that we were approaching – and, of course, the mindset of the team behind the company, their culture, their values, and really, the way they behave is very important for us. We wanted to make sure that should we acquire a company, we could merge the teams and be successful. It’s one of the hard things to do and I believe there are so many opportunities to mess up that we wanted to make sure we had a good understanding of who was behind Morphl. And I’m sure you know it by now, but we are super happy by hearing about all of you – Alexandra, yourself, but also the two engineers that have been working with you. And I think that was already the main checkbox or one of the main checkboxes that we wanted to tick, and you guys ticked that successfully. And then there was, you know, the product and the vision. And as you said, we had similar visions. On our end, we really wanted to make sure that we would be able to build multiple models, leading to multiple high-value features for customers in a very easy way. And so far, we didn’t really nail this ML ops, and all layers required to be very efficient at releasing models. And what we discovered under the hood at Morphl was that one of the big focuses of your team was really about building a framework in order to make these iterations very useful. This was definitely very much aligned with what we are looking for. And the last thing I would say – I won’t speak about all the other discussions that we had with all the departments – from a product engineering point of view, the one thing that we were also very happy to hear about is this affinity that you have with the developer. As I shared, being a SaaS company selling an API, selling a service to developers is not something everyone is doing right. You need very precise marketing, messaging, and positioning to help developers understand what you are building and how it works and what’s going to be adding value to your service. And what we have been hearing from your communications, and your website was definitely aligned with this, and we thought that it was perfect for us. We have always been focused very much on the developer experience. A few months later, here we are.
Ciprian Borodescu: Excellent. And I can also share – and you know this – that I was in a different state of mind, and I had to get back to my own personal why to be able to make a decision for me and for the team. But I’m happy to share it with everybody today, which was that I always believed that the impact you have on others is the most valuable currency there is. And more to the point – and this is something that I also shared in our first all-hands meeting with the entire Algolia team – I and we at Morphl love helping people succeed through the use of AI. And it’s that simple. And by joining forces with Algolia, this is no longer just a mantra, it’s a realization. And I see the Algolia AI products that we’ll launch in the future as a continuation of Morphl, and this makes all of us here beyond motivated to succeed. And maybe we can dive deeper into the product vision and what people should expect in the months and years ahead from Algolia when it comes to AI.
Sylvain Utard: Yeah. And again, we’re gonna share way more news in the next few weeks, but really, what we want to do here is bringing everything that our customers need to build their digital experience. And so, of course, search is at the core of this. And when you go especially to e-commerce or media websites and applications, the search bar is the main component of the website in so many of them. A lot of companies understood that very early in their journey. Of course, I think we’re usually mentioning Amazon here, but they are just only one of them. And the reality is that while ‘search’ is the main component of this digital experience for them, there is so much more that is around. And we strongly believe that with all that Morphl has been working on over the past three years, and with your expertise on AI and on e-commerce, this is gonna fit very well in this vision of building more APIs and more features for developers to build their digital experience, wherever the end-users are. We mentioned it at the very beginning, but I think it’s important. At the core, we have an API and then we’re providing tools for developers to integrate our service across devices across channels. So the idea would be to really provide the same experience on mobile, on your laptop, maybe in different versions, on the iPad, on the tablets. And we even have ways to connect on your Apple Watch, on Apple TV, or, you know, everything running Android. So, we strongly believe that the world is going there where you need to be able to deliver a service that can be addressed and accessed through different channels.
Ciprian Borodescu: So, Algolia has now a research and development center or office in Paris, and through this acquisition, another one in Bucharest, Romania. And personally, I’m super excited about our expansion plans and can’t wait to make AI happen together. And you know what? One thing that we actually didn’t talk about, but I realized that we almost took it for granted is the fact that we did this entire deal online over Zoom, and we never met in person, which is a shame, and can’t wait to meet all of you in Paris or wherever in the world for that matter, even in Bucharest. Cool. So, last year, I had a special section on the podcast, and I think I should keep that this year as well. It’s called lightning questions and answers, a series of short, fun questions that you need to answer really, really fast. Ready?
Sylvain Utard: Okay.
Ciprian Borodescu: Okay. This is a surprise for you, too. But here we go. So, favorite movie.
Sylvain Utard: I really like The Green Mile. It’s not the funniest movie but it impacted a lot my life when I saw it. More recently, and with a more fun edge, The Wolf of Wall Street has really been one of the movies that I keep enjoying when I watch it again. That’s a bit crazy.
Ciprian Borodescu: Yeah, it’s crazy. Cats or dogs?
Sylvain Utard: Dogs. Definitely, dogs.
Ciprian Borodescu: Okay. You hesitated there for a moment.
Sylvain Utard: Yeah, but no. Dogs.
Ciprian Borodescu: Okay. Nice. Favorite book.
Sylvain Utard: I’m actually a very, very bad reader and I keep falling asleep when I read books. However, ask me at any time of the day what’s on the front page of Hacker News, and I will tell you.
Ciprian Borodescu: Okay. Okay. Now, it’s a tricky one. YC or Techstars.
Sylvain Utard: Of course, we have been through YC so my heart is there, but I’ve been a mentor for Techstars Paris for the past two years and I’m really happy to work with and see a batch of companies.
Ciprian Borodescu: And the bonus question. After interacting with you during the last few months, I could tell that you’re passionate about leadership – as am I – leading teams and helping people discovering and surfacing their potential. The question is, who was or is your leader role model, if you have one?
Sylvain Utard: Wow, that’s a tough one. Over the past few years, within the visibility of Algolia, we have been going through personality tests and all these online tests that help you understand where to improve or what to continue to focus on. One thing that resonates a lot through all of these is that people that I’m working with are usually describing me as a team player. And indeed, you know, when it’s about wearing a hat to represent the team, or being the captain of the team, I’m usually the one volunteering, like helping and leading, that’s true. When it comes to who I look after or who my role model is, oftentimes, in my early career, I was mentioning Steve Jobs as the one that really knew what to do, where to go, and really didn’t necessarily listen to feedback from some people because he knew very well where he had to go. And I realized today that this is really not who I am. So, I’m not sure. I’m not sure I have a role model as of today, but I like playing team.
Ciprian Borodescu: Nice, nice. Well, Sylvain, it was a pleasure to have you on this podcast. Thank you so much for sharing your wisdom with me and with us. How can people reach out to you for ideas and comments?
Sylvain Utard: Yeah. Thank you so much for having me. If people listening to this podcast want to reach out, I think the easiest way would be through LinkedIn. I’m usually quite responsive there. So, if you go to linkedin.com, and search for my name, you shouldn’t have a hard time finding me.
Ciprian Borodescu: Excellent. Thank you so much, Sylvain.
Ciprian Borodescu: You’re welcome.