In our conversation, Christophe explains exactly what those early investment days looked like, the struggles of chasing your dream without millions of dollars to lean on, and being thrown into competition with one of the most popular wiki products in the world. We also get to hear what went into creating Ask, Slite’s built-in AI assistant trying to help every team find information faster. Christopher’s enthusiasm and excitement when it comes to what the future of work might look like is so inspiring and I hope you enjoy it as much as I did. Outro script
Who's our guest
Today, we speak to Christophe Pasquier, the CEO and Co-Founder of Slite, a modern knowledge base that helps you escape the chaos of information overload. In 2016, Christophe founded Slite and joined Y Combinator's Winter 2018 investment batch. Since then they have raised more than 15M$ with incredible investors such as Spark Capital and Index Ventures, to serve tens of thousands of teams every month and scale a remote-first team spread across twelve countries.
Max: Thank you so much for spending some time with me. I'm very excited to hear about Slite's story and your story a bit. So if we could start off and you could just take a few minutes to take me through your career in founding companies, your career with Slite and how you've ended up in this position you are in running this team of 40 people
Christopher: We're lucky enough to have closed our seed round before even reaching VO? We had just a massive opportunity and we took it. So we are kind of stress free on that, and we had like a very good growth during YC?, but Notion, which is our biggest competitor launched during YC? and Notion actually had seven years of work before this moment. So it was 2018\. So they had six years of work and we kind of never saw them that much in our discussions or whatever, it was just in existence. The big kind of parting shift that they brought was the mix of database in the text digital world, and that was something massive. We felt that this was stupid. We felt, and we still feel that project management shouldn't have its room into this world, and what we realized in the following years, just by the mere, obvious product market feed of this kind of mix, was that having database in a document environment was not about project management at all. That's still the direction they're taking and that's still the direction that we see our larger customers really having a problem with. And when I say large, just 50 plus people, it just doesn't scale. But database in edit, actually lets you structure a lot of stuff in very smart ways. It actually was a game changer for the core position that we had, which was the single source of truth of your company. Having all your knowledge in one place means meetings and project and processes and postmortem, like everything in one place. You need to have a form factor that suits all of this different kind of data. You don't want the same place towards to handbook. You want to have a nice digitalized book, you want to have the cover and the summary and something that is made for you to browse in a nice and natural way. When you are in meetings, you don't care about the 10, you know, last one. Like you don't get comfortable removing it, but you don't want to see it in the sidebar and you don't want to see that much in your view, in the interface that you are seeing at every given time. That was in inside Lake? there. It's a simple design pattern, but it was a massive shift of parting for our industry. And so for the next three years we were kind of digesting that, growing our customer base. We never stopped growing, of course, and just catching up. We had a one year and a half old product and we had so many features that were not there, like run sharing and commenting and notification and whatever, in mobile. Really just to develop expectation of our market was so high that we needed to just catch up on that info for the next three years really. I wish I was lucid about that at the time. But that was the game, we were just catching up. We had to make the best and to keep our simplicity, to keep what made us unique. That was the next phase of Slite, and the last phase, which brought us to today, is we reached this moment where we're like, okay, we see that for many teams who have the best product, but marketing-wise, it's out to sell or it's out to make it clear. We had two moments on that. The first one was with discussions and our focus on remote and async, when Covid hit. We have been remote since the beginning of Slite, and we saw that some teams that we loved the most by unique connection and the one that we thought were the pioneers were using Slite as a walk tool, as a way to think and collaborate in teams. They were creating a doc, answering to its with sub documents where you will elaborate stuff and so on. The phone factor was not the best, and so that's where we built discussion. And what we realized... So discussion is just inside your knowledge base. It's a place for you to have elaborate discussions, elaborate long form communication that is asynchronous, with a clear, very simple racy framework. Who's responsible of the discussion, who can close it, who should be participating and so on. And more importantly, we saw that this was massive for teams to actually discuss and decide and log their decisions inside their knowledge base, which is basically where decisions should live. For instance, I was with the team yesterday in call, a lead that is starting to use Slite and their pain was that. We have Slack that we use so intensively, just to discuss stuff and I have to look through hundreds of thread to find what we decided. And on top of that, it's probably not synthesized in one message. So this is just not the right way. If you want to work in async, remote and just in any team really. If you want to have your teammates having access to information at any given time, that's just not the way to it. So that was a big bet. We just say, let's really target the pioneers, the ones that are really, really ahead of the curve. Because tomorrow they will be the one sharing out work in the future and so on. I still believe that this was a great call business-wise. It just takes time. It's out to convince people when they come from knowledge base. Well, at least on the concept, it's easy to convince them, but in habits, actually changing your communication habits and form factor, it's actually quite odd. So yeah, that was the last phase, the third phase of Slite. And right now we are really entering a new one, since the last few months, which is really like we've kind of focused back on the core value prop? of our users, which is most of the time an evolution of the knowledge job to be done. Discussion completely fits this, but you know you have people that have questions and you want a tool to give them answers. That's as simple as that.
Max: Exactly that. That's amazing. That's a really interesting timeline of how things have worked and how you spotted holes in the market and positioned your product, not necessarily against competition, but above competition and tried to find your own edge...
Christopher: Yeah, absolutely.
Max: ...to beat people out. And one thing you talk about in your blog - I think it's when you launched discussions - that I found really interesting was you talk about how we work in a different place than we meet a lot. And that's what a lot of companies do is that they meet in one product, in one location and they work in other places, and it kind of completely disconnects meetings from the work we do in actually making decisions. So I was curious if you had any recommendations or advice for young companies to make faster decisions and avoid those silos from happening.
Christopher: That's super interesting. Yeah I think... I'm not sure our way of working would be suited to very early stage companies.
Max: Right, of course
Christopher: For a very early stage company, honestly, I think the best is just to be in call everyday, with whoever you are deciding with, right? Just having a sounding board with your co-founder or whatever, and every day just not accepting if the thing is not shipped or decided upon. For later stage, when you start to kind of feel that you want other people, like certain dedicated people to be involved, I personally feel that important discussions or decisions are better done by writing. I think you have two phases, you have the phase of thinking, ideating, innovating and so on. It's kind of like, are you familiar with the double diamond process?
Max: I am a little bit, but I'd love for you to explain it for anyone who isn't
Christopher: So it's very simple, right? Like you imagine two diamonds, one next to the other. It's a design that is supposed to be used in the design world. The idea is if you are looking for a solution to a problem, you want to not get too constrained and you don't want to go fast to the solution or you will skip alternative worlds that you just haven't thought about. And if I remember correctly, the idea is that you have two similar phases, one on the problem, one on the solution, where you start by expanding on the problem and really speaking about, broadly opening up this expansion part of the first diamond, opening up on what the problem is, what the problem that we should solve is, what really matters. And then closing down on, refining, okay, this is the one problem that we really think is important to tackle right now. And then the second diamond is the same on solution, right? You just go crazy, and then you restrict to the one solution that feels like it's solved the problem in the best possible way. For decisions, to me it's the same, right? You need to have all decisions or whatever - well, I guess decision suits any type of working in a startup - but you just open up by being in the open, bring up a topic. If you say for instance, we have a classic - I think - topic for any startup should be their business model. Just understanding their product market, channel model fit, how all this connects. Should you have set up a premium for the market that you're targeting. Is your price connecting with whatever market you're having. This is the kind of thing where you can't have one meeting, one decision. Or if you do, it means that basically you don't even need the meeting. You already know. So to me it's being clear, actually being lucid and explicit, like "we are in the ideation phase", like "we are in the opening up phase". Just drop ideas what needs to be in this ocean of information for us to then make the right decision. That's to me the phase that happens the best in life. Or you know, like a mix of Flive?, quite sync, like just Slack and so on.
And then formalizing it into long form with a clear question, clear decision maker. I think that's for larger companies, anything above 20 I'd say it's a very, very good process.
Max: Yeah I think that's something that I've definitely learned from working in companies of varying sizes. I've worked in six hundred person companies, a hundred person companies, and now Claap. I think we're 14 now, so still in that early stage. And something that happened three years ago when everyone started being remote - obviously you guys were remote already - but when a lot of companies had to adapt to remote models, what people are still figuring out is how to create processes that don't just drown out the amount of production you can do. I think that's an easy trap for companies to fall in is to just create too many processes, and create too big a tool stack. So is that something that Slite tries to address, or tries, or thinks?
Christopher: I mean it's interesting. I totally agree with that. I don't think it's all responsibility in a way. When we educate about this, we definitely have an article on this, on killing the process. I think it's oversimplifying to say that processes are just evil, but there is some truth to it. I have on my bed table right there, the Netflix book: No Rules Rules, which in TLDR? is the Netflix founder saying - I don't know if he founded it or he was the CEO of a company and then started Netflix and tried to not replicate the AI made in the first one - And basically it was whenever a problem happens, 1% of the time adding a layer of process to make sure it didn't happen in the future. Basically doing that makes your company doomed to be bloated and just not operate anymore. So actually whenever we have a process that comes in - that's something classic from HR - I'm pushing back. I am kind of a pain in the ass for them. But in terms of what we can do as a product, I think as long as we are simple enough, there are UX decisions to not make it so that we don't block things. For instance, something that we get asked all the time is approval workflows. When you want to edit the documents, it's still something that we think about, right? But it'll never be the default. And that's what people that use Kit Up or Kit Lab?, some technical tools to handle their handbooks are used to. Having a way where you edit a whole document and then have to understand technical terms. Doing a pool/poo? request and asking for someone to merge it, which is basically like "change this, here is a new paper, can we change this in the Bible? And that's not the philosophy. That's not the default philosophy of Slite. We want it to be open-ended by default, so that anybody has no barrier, by default once again, to edit. I think that's the best we can do.
Max: Yeah, exactly. It's like you say, your frustrations with Google Docs. I think a lot of when I get frustrated with it comes from that double edge sword of having much less privacy on a document. Having everyone kind of equally able to edit documents sometimes creates more confusion and more processes and comments get resolved when they aren't resolved. It's very easy to lose track of the decisions you've made. Something you've spoken about before is the rate of iteration about how young companies can really, really find success easier is how fast are you iterating? And you can only iterate if you are making solid decisions, right? So I was curious how you guys go about your decision making.
Christopher: Well, that's interesting. I would argue that... Yes, you're right. If you mean solid as in being extremely assertive, you're right. You should basically be comfortable being wrong, right? One mistake that we did, we had this big moment of freezing for one part of Slite for six months on discussion, actually. We launched it and then for six months we saw that activation was not big enough, that people didn't get used to it enough, at least not to deliver of expectation that we had. And we kind of discussed a lot with the squad. We need to push it further. People don't see it, don't understand why it connects with knowledge base. They need to be sold on the idea that your decisions ease your knowledge and if you don't get it and if you don't embrace this idea, then of course people won't activate. But it was just not pushed both in product marketing and in the onboarding flow and people are like this button called 'discussions'. But they add to figure it out by themselves, and we had so many discussions in one way and the other. So we ended up after a few months, the squad didn't change. They improved the feature inside, but the whole activation part, which was the big problem at the time, wasn't touched. And so that's a good example of a bad rate of iteration where we should have just accepted to make a bad decision and found the ability to roll back after. To me, decision making is just about the asset of the team. I think every founder has that in them. One mistake that I did, I think for a long part of Slite was being afraid of my weight. So, for instance, like a good example of that is there was this podcast with Wade Forster, Zapier? CEO, explaining a system called the hashtag system. So the idea was, he has so much weight into decisions, that when you say something, people will just change completely their priority and just say whatever he is saying. So he said I was bored of it because, I have a strong opinion. Sometimes I want to share them. I think they are valuable. But, you know, this educates me to just not say anything anymore. So we put this system together, adding a hashtag to whatever message you will send. That will be a feedback or challenge saying hashtag 'strong', hashtag 'mailed', hashtag 'FYI'. Kind of saying, 'FYI' is just I really don't care, I don't have a reply to this, but I had this idea the other day or whatever, hashtag 'strong', or you have the hashtag 'implo', which is basically stop the discussion, I want this done. I have like all the information in my head. I know that this is essential. This is the maximal level that I can ask you that. And someone probably never uses it, but when I read that I was like, yeah. That's so true. I would hate to have a CEO that will be an asshole just telling me what to do. So for a very long part of a Slite, I actually tune myself down so much. Not making tough decisions. So for the first two years you don't have a choice. You're creating everything and you are so small that it's fine, but when you start to reach a certain point, you start to bias yourself. You're like, okay, we have 20 people, 25 people, not everybody's at the same level of information. And in my case at least, I was like let's not be a CEO asshole, and so let's not give too much opinion, and I was actually not sharing enough and I was not tough enough, and setting the right level of standards. So if anything, I think you know how to make good decisions. In my case, I would say just accept that you need to make decisions basically. There is some form of education in the team to say as long as you are fully an owner, you take responsibility if things go wrong. The person taking the decision is a blessing for any team. So I think it's a form of internal education really
Max: Yeah, and that communicating internally, especially you guys know as you've been remote the entire time and you work a lot async, you know exactly how these decisions can sometimes take a while, and you have to proactively make decisions happen. They don't resolve themselves. I was curious, because I know you are a big async advocate as well, I was curious what the meeting policy is at Slite and how it's changed over the years as well.
Christopher: It's a good question. So we don't have a meeting policy...
Max: Perfect. That's good.
Christopher: Yeah. But I think it comes back to what you said about processes. We try to reduce that, but it's the same as values and principles. We decided to not have values, but to have principles. To say, it's not like values. It feels like you never can cross. It's moral, right? It's a form of ethics. You can't cross the border, the line.
Yes it feels more like commandments and something.
Christopher: Exactly. And principle is like, in principle, you're supposed to do that, but if you change this one or two times, that's fine. And to me, process and culture is the same. Process is like, this is the way we do meetings and so what we give in meetings is how to communicate guide at site and basically we say we don't believe in a remote, especially that you need large meetings to get you done. The only places where we think meetings are useful is for human connection. Or if you are less than three people, for ideation and creativity sessions and so on. That's the one bit where deep work doesn't work well, is getting excited and it's getting creative. We need human connection for that. So that's the way it shows in our team. Over the years, I've developed a bit to take a weekly meeting, which we didn't have a few years ago. That's something that I think changed. I had my co-founder sharing a screenshot the other day in Slack. You can see people in a huddle like the call feature of Slack, which is something that we use very intensively. He was showing a lot of people in huddles and he was saying that's amazing. We are async, but it's really important that people feel comfortable just jumping the call with each other and so on. To me that's the balance that I love is big decisions there are in async. So by writing formats and so on, the work day to day. We still call to have these creative moments, these human interaction to crack jokes, to have fun. So no policy, but that's the way things work.
Max: Good, good. Yeah, I was gonna say, you know Claap. Our whole policy is we wanna help people work async, but something we never, ever want people to think is that meetings are dead or we can't have meetings because, as you said, you know from being an early stage startup, it's about being in calls every day basically. And the huddle feature in Slack is actually something me and Pierre, the co-founder, we use every single day. Because it really brings the gate down between you and meeting someone. Sometimes putting a meeting in your calendar can feel like you just want the meeting now. So something like huddle where it's like, hey, I need you now, is so exciting.
Christopher: Yeah I love it. Michael and I, basically I had to cut down... We still have very frequent ones with the main peers. But I absolutely agree. You can feel that people coming from other companies typically - I had this with Vita, my head of brand marketing yesterday - all week limiting for the leadership team and we're having to discuss something together. And I was like, I'd love to pick your brain on this. We found the time and we were like, okay, after the leadership meeting, great. She sent me a calendar invitation and I was like, why? Literally there is no way nobody can book me. People don't book each other like at Slite. So that was interesting, a cultural shift or whatever.
Max: It's about keeping a log of information and I guess that's something that Slite, your whole product is a log of information. It's a database. Speaking of your product, and something I know is very new and exciting to you guys, and a big topic is Ask, your new AI tool, which helps people. I know Notion fear is a big fear at companies I've worked at where you're looking through Notion. When I've worked with Confluence I had the same thing. I'm like, this is so much information and the search bar maybe actually doesn't work the way I want it to. Something you've spoken about is the bright side and the dark side of AI, and I would love to hear that explained just a little bit and how you've managed to put Ask into the bright side of that argument.
Yeah for sure. I mean, it's so exciting and Ask is just one part of it. It happened in a very unexpected way. So the dark side and the bright side is really, Air exploded. We had played with LAM, so large language model for beats. We had this week of ideation and prototyping - very good example of when you need async - So we were mostly in calls with Flo, one our lead developers when everybody was off during Christmas holidays. We had this target list of stuff that we wanted to try. Can we do X and Y? So the dark side is obviously intense, but the side where we think there is maybe less interest in generating content inside your editor. So Chat GPT and GPT overall is amazing technology, but it generates content for you. It doesn't make you think. It lets you reflect maybe, but all the stuff that it creates, like it's ? for your teammates. So to us the big risk, it's amazing. You say, create a table or whatever. I mean the examples are purely like why? But what do you use it for in real life? You say, create me a database with content article about remotes and tag, who I should speak about on this topic and blah, blah, blah. It does this thing, you ask the question, what's your role? Of course if it was perfect, that would be great. At the same time, it's just a first draft. You see it and you're like, well I have to redo everything because I don't think this article ID is that interesting. And of course, GPT found it, it means that it learns that this will be a natural completion of this kind of prompt. So it means that it has been spoken in some way or another in the past. And so all the unique ideas that you can have, they can't appear through GPT. So that's the whole thing about generating text, it generates junk, and I think will be mistakenly thinking that it replaces your work or your thinking. We had this exchange - sorry I get in the old direction because I'm so passionate about that - but we had this exchange with people on LinkedIn yesterday. In a way, GPT and I might teach us to detect what's human. Like the little queerness that makes our writing us. Like I write in a weird way with a lot of exclamation points or whatever, with maybe a lot of French little translations.
Christopher: And you know, GPT will have time to do that, but also it won't be perfect. And that's what we are, right? So in a way, I love this thing because if we manage to learn to detect what Chat GPT can do, it actually can let us double down on what makes this unique. So anyway, that's more a philosophical point. So we thought in the context of knowledge base and documentation tool, that if you start using GPT at scale, you will just create data that will put you the search and the cyber of everybody in your workspace, and that's something that is extremely dangerous for us. A good example of that at lower scale is what Notion did with their web lever. It's a great growth tool. So you can have a Notion extension and clip any page, or you can have API integration and just bring tweets to a database, this kind of thing. People think they love it. The problem is that these folks, they search and Notion search is known to be very, very bad. And one of the reasons is if you bring some pollution, of course it'll be bad. It's hard for a system to understand what's junk and what's not. If you cannot promote junk creation, you can't complain that the thing doesn't work when you know your team is trying to look for answers. So anyway, that's the dark side.
Max: That's the dark side. That's a very interesting point. I think everyone's debating that topic at the moment, and I think everyone's also trying to find a way to bring AI into their tech stack because there are, like you say, good can't exist without evil. Humans can't exist without AI. So I like the point you make that it makes it much clearer what's written by a good writer. Something like Chat GPT or something, anything, any language learning model.
Christopher: Yeah, yeah, a hundred percent. Probably you are a writer, right?
Max: Yeah, yeah. A lot of what I do is writing.
Christopher: So I'm thinking about Mel our content writer. I'm thinking about other people that I know in the industry. Mark that used to work with us, who was working with Typeform, amazing writer as well. And I see people being a writer...
Max: Is that Mark Delevy?
Christopher: Mark Sinani
Max: Oh, Mark Sinani, okay. I'm ex-Typeform, so I recognize the name. There's too many talented people coming out of there.
Exactly. But I think his writing is so unique. It'll push us in our limits. We let get people to not do SEO shit? Actually they're right. SEO is machine speaking to a machine. Should I rank this up? Nobody ever enjoyed writing SEO content, right? Right now we are in era where you want to create genuine content. Anyway, and so in the case of the bright side, we played with it. I think by the way, you probably have a use case for Claap as well. We played with GPT and digital. Kind of interesting. We felt it was polluting already. We felt it shouldn't be in the editor. It should be on the side, helping you write maybe, but it shouldn't be auto completing in the editor. So that's something that we keep exploring and then we played with it on discussion, which was absolutely amazing and we definitely would shape something around that. Helping you summarize long thread of information helping you understand the blockers. Who helps and ho wants the decision to move forward. Who wants to block it, understand the arguments. Maybe even understand, is it really being objective here? Is there emotion? So you can really think about stuff, and I'm sure Claap can do it as well. Really helping us collaborate. So that was super exciting. So the last one was Ask, and Ask was the most difficult tool that we could build, and we were lucky with Flo. This is our head of security, but he is a very talented engineer and he got so excited about that, and I would never have been able to think that we were capable of building that so fast. But it just pushed and pushed and pushed. In a nutshell, Ask is an assistant. It's like your team library, but a very smart one, kind of like a robot assistant, really. So imagine somebody that has read all your content. You've taken a meeting this morning, this person has read it, and you can come to them and just ask whatever question and they will look through all the books of your company and say well, here's written that you know, your company vacation policy is 16 weeks long and so on. But here's the answer and here is the source. So Ask is really more than search. It's really a new way to unlock the content of your team where you just literally type and enter, you'd find the sources everywhere in your workspace, give you the answer back and explain. You can highlight the sentences and we'll give you what we think the paragraph of tex that we think created this sensor. So it's really a way for you to get the answer but also fact check it straight away. We see it as the keystone of the next generation of knowledge base. It's the perfect way for your team to ask questions, flag when the question is not documented, because sometimes we can't answer. So that's the perfect way for you to open new questions and say well, I wanted to know what's the process to provision? your article on the blog maybe, and this has not been documented yet, and just tell me that. I wanted to know what was the pricing of the new feature that we are going to ship next month and maybe that's not been decided yet. So great, let's just open it and we'll like it at the end.
Max: That's so good. I can think of multiple use cases within the past 24 hours where I would need that
Christopher: Do you have them in mind? For real?
Max: Yeah literally right now I can think about yesterday, I was creating a video and I was trying to find our logo, and because we're a young company, the logo has changed a lot. I wanted to know, hey what was the logo a year ago? What was the company logo six months ago?
Christopher: Wow yeah, that would be amazing.
Max: Things like that. Whereas I was like, okay, well great if I can also hunt through assets, and something you've spoken about a lot is assets expiring and what you do with expired documentation and expired assets, and I think if you can combine those things together hat's so incredibly useful.
Christopher: The assets one is super interesting. We never thought that much about files, but that makes a lot of sense. We do have brand guidelines where we reference files, so what you said makes a lot of sense. Yeah that's very, very good use.
Max: Cool. We can look into it. Amazing. I've got one last question to ask you. I've already learned so much, but I always ask everyone this last question just to finish. If you're not the kind of person who makes predictions that's fine, but I would love to hear what your thoughts are on, in five years what's one process that you think will 100% last the test of time. Or if there's one new trend in product or in the startup world or tech world that you think, wow, this is gonna change the way we work over the next five years, and it can be something we've talked about.
Christopher: Yeah I mean, obviously AI. Right now, I think there are other stuff that could be ballsy. Like I would love all startups to have a dedicated long form written way of making decisions. I think everybody will win so much peace and calm and clarity over their work, but I can't say that I can predict it. AI changing dramatically the way we work. Being an assistant to most of the interaction we have is absolutely obvious to me.
Max: Yeah, of course.
Christopher: It's not even a prediction at this time at this point, I think
Max: It's just exciting to see what may happen. And something we're gonna do in season two is be checking in on all the companies who we've spoken to about their new AI tool. And by the time this episode comes out, our new AI tool will have come out. I can't tell you what it is yet. You may already or may not already know, but by the time this episode comes out in April, you will have heard about our approach to AI as well.