host intro

Welcome back to the show. I'm John Steele. This is the Funnel Vision podcast, episode seven. And today I'm going to be speaking with Charlotte Schmitt, coming to us from Zurich, Switzerland. Charlotte is the CEO and founder of a company called Octolens.

Octolens is a social listening tool. And they're now setting out to answer the question that seemingly on every marketer's mind. How do we gain visibility when the AI platforms seem to change what they want us to do on a daily basis?  So I'll be excited to dig into that with her a little bit. But I also want to talk to her about how she came to found Octolens because I think the story is really instructive for other founders out there.

Now, this isn't Charlotte's first rodeo. The first year out of business school, she says she went all in on a startup that was the first one she was ever pitched. It was called SaveSpace. And it was a full service storage company with trucks, warehouses, and she even got into Techstars. But two years in, she decided it wasn't really for her. And she decided to go a different direction.

She linked up with crowd.dev, a platform that helps companies build developer communities. While crowd.dev became popular, helping companies measure their own communities, they soon figured out they wanted to help their customers measure what other people were saying about them elsewhere. That's where Charlotte came in. She led the project of the creation of an internal tool called EagleEye. The Linux Foundation bought crowd.dev in 2024. By then, Charlotte had already taken EagleEye to the next level with the founding of Octolens.

Octolens started out as a SaaS platform you're probably picturing in your head right now. A sleek dashboard, a cool app, drag and drop integrations with social media. And they still have some of those things. But an investigation in late 2025 revealed that their largest clients never use the app at all, pulling data into Claude, ChatGPT and Slack instead. So Charlotte deleted the analytics page and rebuilt Octolens to offer the API, webhooks and MCP server on every plan. They call it social listening for the agent era.

Somewhere along the way, she swapped engineers, initially starting the company with crowd.dev co-founder Pepe Riero. When Riero left, she brought on Igor Kottua, a growth engineer from crowd.dev who had since left, founded and sold his own startup. Today, with a team that's just the two of them, they serve hundreds of customers, including names that will be familiar to many engineers, like Railway, Cursor, Vercel and PostHog.

That's not because they haven't gotten around to hiring. Charlotte says they've automated nearly everything that doesn't need a human. Through automations and processes, she says she keeps pushing back the date when she thinks she'll need one.

So I wanted to know, how do you take an idea from your old company and make it your own company? How do just two people serve hundreds of customers? How do you build a company on AI when the model seemed to change daily? And how does she see current solutions to things like AEO, AI citation, and social listening? Let's get into it. 

interview

[02:49] John Steele: Charlotte Schmitt, welcome to the show!

[02:51] Charlotte Schmitt: Thank you. Thanks for having me, John.

[02:53] John Steele: Absolutely. Absolutely. So you're coming to us from Zurich today. Is that right?

[02:57] Charlotte Schmitt: Yes, I'm in Switzerland, close to Zurich, by the lake here.

[03:01] John Steele: And how did you how did you end up working from there? Are you are you from that area or?

[03:06] Charlotte Schmitt: From Germany, so not that far. And now it's already been around six, seven years here in Switzerland. And yeah, it's a it's a nice spot. It's not San Francisco yet, or the startup bubble, I would say, but it has a nice growing tech scene. A lot of the bigger tech or AI companies actually have bases here in Switzerland as well, which there's a big Google campus, etc. And a nice technical university. So it's it's actually a nice budding area for for startups, I would say.

[03:38] John Steele: Cool, cool. So tell us a little bit about Octolens. What is it? How did you come to build it?

[03:44] Charlotte Schmitt: Yeah, sure thing. So I'm one of the founders of Octolens and Octolens is a social listening tool for tech brands. So we help our customers find out whenever their brand is mentioned, essentially everywhere where people talk nowadays. So that's socials could be a podcast. It's just wider web. Let's say you're on a blog or a competitor comparison page news as well. And we let our customers know when that happens so that they can then do what they want with that data or that piece of information. It could be responding to it. It could be just going more broad analytics on that. And yeah, that's a kind of condensed version of what Octolens does.

[04:28] John Steele: And this grew out of your experience working at crowd.dev, right? So you were you were looking to track mentions there. Is that how you got started with this?

[04:36] Charlotte Schmitt: Yes. So before Octolens, I was chief of staff and then head of operations at a startup called crowd.dev and crowd.dev was a community led growth platform for dev tools specifically. And I don't know if you know much about developer tool growth. It's one of those domains where tools or companies tend to grow bottom up. So it's developers discussing a lot on Reddit or X or back then a lot of Stack Overflow. And so the way you grow your tool is getting developers using it and recommending it to each other because they don't do well with like your standard marketing. And they kind of if they hear marketing speak or ads, it's not going to work very well. So it's like this peer to peer recommendations. So community mentions and content is super important. And back then, I remember we were doing research and I was as well looking at social listening tools for us and also for our customers because that question would come up quite a lot.

[05:38] And there wasn't really anything great out there for dev tools and B2B SaaS specifically. And it was around the ChatGPT moment. And we also noticed there was no tool which really used AI yet. So that was a moment where we thought, okay, maybe we can do something around this if we can't find the tooling. And in that crowd.dev platform, we developed a feature called, it was back then called EagleEye. And that was the initial kind of inspiration for what Octolens is today. And that tool got sold to the Linux Foundation. And I just went ahead and started building Octolens from that.

[06:16] John Steele: So this is your second bite at the founder Apple, right? You founded a company that went through Techstars, SaveSpace. Can you tell us a little bit about that experience?

[06:27] Charlotte Schmitt: Yeah, sure. So that's completely different industry, completely different product. That was my first time being a founder. Relatively soon out of business school, I first, I always wanted to try to found something. But I don't know, business school always tends to push you into, you need to do consulting or finance or something. So then I first did a couple years of consulting, but at my first chance of someone coming to me with an idea, I immediately quit, immediately went all in to SaveSpace. And it was a storage, storage startup. So that meant we were rethinking how storage should be done, you know, how people bring their stuff to storage, leave it there for years, don't really have access to it. And then it's outside of the city. And it's a lot of hassle. And we decided we do like a full service approach. I think in the US, you already had companies like this. But in Germany, that wasn't really a thing yet. So we decided to do that. And so that was very logistics heavy. You have trucks, you have warehouses. Yes, there's like a tech aspect, but it was yeah, asset-heavy B2C.

[07:30] And after two years of that, and going through Techstars, I decided it just wasn't my industry. It wasn't the customer that I wanted to spend all my time on. And as you have heard, probably having a startup is like, in a co founder, it's like a marriage, you think of the next 10 years, and I didn't see myself in that space. So my co founder continued on, he's still doing that business, which is great. But I decided, okay, I want to see B2B SaaS and I want to do something completely different. So that's when I then joined crowd.dev in that journey and in software already started. And let me now talk to it.

[08:05] John Steele: So yeah, you mentioned that working at crowd.dev showed you engineers try and find information online and how to reach them is different. They're allergic to marketing speak and things like that. I wanted to ask you about social listening, specifically because the only way I know about social listening is through the marketing side. And a lot of the tools that I know have existed in the past have been through the marketing side. It really doesn't seem like yeah, your your tool is your customer is necessarily marketers or at least not, you know, maybe marketers at tech companies, but also it for people trying to reach marketers, it seems like they're trying to reach or end users or customers, it seems like they're trying to reach developers. Can you talk about the the shift in social listening and why maybe we're seeing a shift from where this was purely a marketing, you know, adventure to now being more on the tech side.

[08:59] Charlotte Schmitt: Yeah, so Octolens we do or we did, especially because of also my background, like you mentioned in DevTools are with working with developers. We also got our start with most of our customers being DevTools. If you also go on our website now, you'll see a lot of like our customer stories and highlights or developer focused companies. We have ones like Vercel, Cursor, there's Railway, a bunch of DevTool focused companies. And that's a bit because of my background and also the platforms we covered. So we definitely cover a lot of the platforms that are important for these types of companies. So that is places like Reddit, it's X where opinions are really formed on products also very fast. You have to be in the conversation and on it very quickly. So that's why we serve that segment very well. And when I and as I mentioned the tools that existed before, we're very are very much still also focused on marketing teams and maybe some dashboards within the app that show you like trends over time, sentiment, etc.

[10:04] Which is still great and very important. But our users also being very technical companies often, not exclusively, but it is a large segment. It's a lot about them having access to all of the data. So we made sure always to allow our users to do to own the data that we serve them. So that means also webhooks, getting data out, slack alerts, CSV downloads. Now it's important, you know, to have an MCP to connect to Claude, a read and write API. So there's so much interesting data on socials and via tracking specific keywords, your brand, but also competitor and industry terms. The information that comes in is really important, yes, to the marketing team, but almost any of the teams in a company. So, for example, a product team gets really useful feedback on socials nowadays, community advocates or developer advocates. They want to know exactly when the brand is mentioned to go help answer questions.

[11:05] So it's really a tool for all teams beyond just marketing nowadays, because so much is decided on socials nowadays as well.

[11:17] John Steele: Yeah, so I think it's interesting that the tools that you cover, Reddit, Hacker News, LinkedIn, I believe was in the list there, GitHub. Yeah, these are definitely where the developer discussion is happening. My understanding of social listening before was sort of the goal of it was to understand the sentiment of your customers towards your product and sort of understanding how they're using it and the type of buyer that you're looking for. In this case, it's much more of a B2B use case, and it's much more of you kind of know the type of buyer to a certain extent a little bit better. Does that allow you to do different things than a marketing instance of social listening might? Where you know that the people you're trying to listen to are developers and potential customers of the product or service that somebody is using Octolens for. So does that give you sort of a different pathway to approach this?

[12:17] Charlotte Schmitt: Yeah, so I would say now we've grown. We've really outgrown just that segment. So our users really span beyond just DevTools. There's you know, AI companies, just regular SaaS and even some B2C now, because we also cover platforms, you know, like YouTube, TikTok, Instagram over time. But I think many startups and us included, we really benefited in the beginning to have quite a specific or more niche target. Like the more you niche down, the more, of course, you can build perfectly for a specific user. And we definitely got a lot of traction in the beginning because we took this B2B SaaS DevTool approach. And for this type of user, it's so important that they can do with the data what they want. So as I mentioned, getting the data out, getting having a webhook because we might have a native Slack integration, but they want to send the data into their data warehouse, build their own tooling on top of it, which is which really helped us.

[13:18] Also, because that's also nowadays how almost all users want to use a product. So back then, it was the more technical audience that knew, OK, I want to get that data and I want to build my own internal internal tooling on top of it. And now with obviously almost everyone being a developer with Claude and ChatGPT, you have go to market engineers, you have marketing people that now are acting similarly and, you know, a software company that's our customer. So for them, it's so important that we have an MCP that we offer an API so that they can build the tooling that they want inside their company. So I think it really gave us an edge that in the beginning, yes, we could market to a specific, more technical company and buyer. And now it's benefited us today because we have all the features that now almost everyone wants because people are all more technical and want to have and build their own tools on top of their social listening data. So that's been yeah, it's been helpful in our growth and we got a little bit lucky in that sense.

[14:21] John Steele: Yeah, I understand from the standpoint of someone building with a tool like Octolens that they would want sort of something that was more extensible, you know, more diffuse, something that they could they could sort of build into the flow of what they're already building. I'm curious, though, you know, have you this is a second iteration. You guys had a sort of a dashboard app type product and you've since moved away from that. Can you talk about that decision for you as a company? So, you know, the concern might be that, you know, the use case is a little bit more loose and sort of less sort of specific. And so that might make things harder for you guys to sort of differentiate yourselves. But also, you're pitching to developers. And so I understand how how required that might be. Can you talk a little bit about that decision and weighing that? Yeah.

[15:15] Charlotte Schmitt: So we we still have a dashboard and we still have the way that users will first sign up to Octolens is even though we try to make it as easy and kind of feel magical as possible with our onboarding flow. We still have that first initial moment where a user will come to our site and then set up the workspace once and then it kind of very easily now splits. We have a large part of our users who then essentially don't log in again and do everything via Claude or ChatGPT or they just use our product in Slack. And then we have some users who still enjoy this going to the dashboard, checking their mentions, they're looking at the graph in Octolens. But we did cut down what we have in the UX quite a bit. So we used to have much more built out analytics page and we just decided at the beginning of this year to delete that because everyone has their own needs of what they want. This user wants to share of voice graph.

[16:16] This person wants something to do more on like trending topics. And so this is what people like custom built in their agent of choice. And just from August, sorry, from September to October, MCP usage and API usage 5xed again in one month. So the trend for us is very clear that our best users or power users are doing so much more from their agent and not forgetting about the ones who are still in the dashboard. Sometimes I do for you do get into this little bit of a bubble. I think everyone knows what an MCP is and wants to use Claude. But then it's not everyone. It is that direction, though. So we just made that decision to cut down the features inside the app and focus on making the data really usable outside of it. And once you have users building their workflows on your data, it's also a very nice sticky product because once you're part of that stack, they're also going to usually stay for quite a while. But what you addressed is a difficult part of this is that once users are doing whatever they want in their agent, it's also a little bit less clear what your.

[17:23] Couple of main use cases are and to market, you become very easily into this like, oh, you could do anything bucket. And that's also something to be aware of because then, you know, your marketing can become a little unclear and you still need to remember kind of what the main use cases are that your users in the end want to achieve or the results. So it's a balancing act, just being the data layer, but then also having an opinion on what are the main results that users get with with your product.

[17:53] John Steele: Let's talk about social listening a little bit as a practice. So what what do you think companies should be listening for right now?

[18:01] Charlotte Schmitt: Yeah, so I always when I give my demos of Octolens or when users ask, like, how do we start? What should we track? And what are the main things that we can get started with? I'd say you usually track a combination of three things. So you'll track your brand name, you'll track some competitors and you'll track some industry terms. And the first bucket is quite clear your brand name and versions of it because you want to know what am I being mentioned? Is it positive? Is it negative? Do I have to go answer a question? So that one is always relatively clear to most people. Then there's another bucket of competitors because you can learn so much from what are they releasing? What are users liking, disliking about your direct competition? And then thirdly, you have this industry, what I call industry term bucket. So for me would be things like social listening, keyword tracking, brand monitoring, because there you can find a lot of great data as well on what are people looking for? What are some content ideas that you could write about?

[19:03] Sometimes you get a lucky strike and somebody is literally asking on Reddit, what is the best social listening tool? And you can go in as a founder and present yours. So that can even be like a conversion or buy intent event, which is great. But those are the three things most companies should be tracking. And from that, you can you should interact a lot with your community and know which threads on those platforms you should be engaging in. And nowadays being very active is almost non-negotiable and will also dictate a lot how you're showing up on search and how you're showing up on AI search. So it has a lot of benefits for internally the company, the product, the content, etc. But then also on the growth side, visibility and AI search visibility is a big one now that social listening helps companies.

[20:00] John Steele: So you've asked people publicly to sort of build things with Octolens and share them with you. Have you gotten any feedback on that? What have people built? Sort of what have they told you about that experience?

[20:12] Charlotte Schmitt: Yeah, so we have I think one of the more common ones now is you have this new new role in the last few years as a go to market engineer. So often or several of our customers have built their internal kind of signal tool where they like to combine all types of signals across different platforms. And one might be socials. So then they'll pull in all their social signals into their overarching go to market internal tooling and use that for doing outreach to companies, understanding who might be interested in their product. So that's one aspect. Others have built their own just dashboards on the analytics that's important to them. So let's say how sentiment is trending on Reddit over time. Other platforms as well share of voice graphs next to competitors topic analysis. So yeah, there's a there's a bunch of different tools our users have built.

[21:12] Railway, for example, one of our customers has also built their own dashboard, of course, deployed on railway PostHog. One of our customers likes to pull social data next to their product data. They do that with another tool called Val Town and they visualize everything in their own platform. So yeah, there's a lot of a lot of interesting tools being built with with social data.

[21:35] John Steele: Yeah. So you can see the similarities, obviously, between, you know, social listening done by marketers and social listening done by sort of founders and engineers that are building products for for developers. Sort of, you know, people complain about certain things or they lack certain features that they want for certain use cases, things like that. And then you try to be responsive to that. Are there any differences between how company who is, you know, making and selling something on like Instagram, maybe as like a B2C product might react differently than maybe like a B2B company that's pitching to developers and engineers. How would they sort of use listening data in a different way?

[22:15] Charlotte Schmitt: Well, yeah, we don't have many B2C products on Octolens or users. So I very much have like the B2B lens. And there you also have quite a big, I would say, difference between like your older traditional companies who really just like to often analyze the data, take that data in and bring it to their different teams versus kind of your newer company like let's say Vercel or a railway. They're being very, they're so on top of everything that's going on online and also doing the replying. So really talking to the end user on let's say X or Reddit, like you'll have founders just going into threads of anyone that just posted about their business to get feedback, to talk to them, which I think is something that companies would have done or do a lot less that are more traditional or old school. And it has like such a big effect, let's say, if the Vercel CEO is personally replying to someone who's complaining about a bill of Vercel and getting that fixed right away.

[23:16] So this very much like talking to users, being within the community and day to day replying, I think is more unique to like the newer type of company we're seeing and social listening versus let's say like a big B2C company. It's more around really like seeing the trends and the analytics aspect of it and building all the dashboards. That's one difference I definitely see just from our users.

[23:42] John Steele: So let's talk a little bit about AEO. So obviously the ability to sort of track and source mentions on Reddit, you know, has a lot of crossover with recommendations that we're seeing around how to gain AI visibility. And that's something that you guys talk about a bit on your site. Obviously being able to track mentions across many different places that are influential that, you know, people are looking at will obviously, you know, sort of increase your visibility. But there's been a lot of volatility in that space. It seems like the pitches kind of change day to day. You can't really tell what is working. What isn't. People are nervous about when they see things like 86 percent drop of Reddit citations in like a four day period. That is a level of volatility where, you know, that their their CEO said like, well, nothing to see here. But I mean, I think there obviously is something to see there in terms of if you're trying to plan your life around something.

[24:43] So talk a little bit about that and how you guys navigate that, how you're seeing the AEO space, what people can trust and sort of what you've come to learn.

[24:52] Charlotte Schmitt: Yeah, it is a very new space still and things changing all the time. At the same time, I think you see a lot of similar things always happening in SEO where one day to the next, Google will make a change in how they are serving results. Like there's always some big splash news that people want to get reactions from the Reddit one specific, I think, is was very ChatGPT focused. But Reddit, I think, is still ranking quite high in, let's say, like Perplexity in Google, AI search overviews. And it is still being used as a source in ChatGPT for sure. But yeah, it can change all the time and which model uses which citations more strongly does shift quite a bit. And I think it shows you that you shouldn't put all your eggs in one basket or focus on one platform. And so Reddit is important. But then YouTube, for example, is also ranking very high on citations. LinkedIn is doing very well in being cited. So it's around really being and raising your awareness of your brand across all these different platforms.

[25:58] And when AI sites you, you tend to do well if it's picking up citations from a lot of different places. And I mean, when you ask AI search something, it basically does a fan out and looks at results from a lot of different prompts and ways and asking the questions. So the more your brand is mentioned across all these different UGC platforms and not only those, you still have your traditional, you know, blog posts, PR, podcasts, et cetera. That's going to help you and also help you stay safe and not just dependent on one platform being cited because, of course, they can make these rather drastic changes from day to day.

[26:33] John Steele: Yeah. So you mentioned sort of what was going on with Reddit and ChatGPT. You know, do you see them sort of settling in on certain sources? You know, I think one of the issues with AEO that a lot of people run into is they're just not sure where and how much emphasis to put on which platforms. And I think being having worked in SEO for many years, I can tell you Google was always a black box to us. We were never 100 percent sure what was going on, what they would say would contradict with what studies had shown. So in your own testing, what are you seeing that sort of has been working for you guys with the sources you do have? And then sort of what do you if you're reading the tea leaves a little bit, like, where do you see it going? Do you think they're going to settle on certain data sources or, you know, give us some stability to work from?

[27:25] Charlotte Schmitt: Well, I mean, I think there already has been quite a bit of stability overall that UGC content is one big bucket. Then just your traditional best practice in SEO for your own site is still very important. And other blogs and high authority places. I mean, I don't I can't know for sure. I'm not sure which platforms I think will continue to spike or I mean, I do. And I did post about this recently. I think one where you or where especially B2B SaaS, I think still has a lot of potential is YouTube. So you have kind of this double benefit that YouTube videos or videos are less are less like attackable by AI. Users are still going to want to watch videos if they want to learn something. And then on top of that, AI will also cite YouTube quite heavily and it'll help you rank very high, especially I think on Google search results. They'll draw a lot from YouTube as well right at the top. So I think one of the more defensible platforms that we're investing a bit more time in now and then I think is more AI safe than let's say others like listicles.

[28:33] I think people aren't really reading those anymore. And they'll just get less and less important over time. So if I'd had to choose, maybe I'd put a bit more emphasis on YouTube and there's still I think in B2B, especially a lot of wide open fields. If you do some searches around your most important keywords, you'll see outdated videos. You'll see a lot of potential to post. So I think maybe that's one. But then again, just not focusing on one platform and making sure you're responsive everywhere, especially where your own users are. So even let's say maybe Reddit is decreasing a bit in importance in AI search. If you know that's where your user hangs out, you should still be putting your time into to Reddit or X, even though X might not be listed as much in AI search. If that's your buyer, if they're on that platform, that's where you should try to be super present in the community and in discussions.

[29:22] John Steele: So, you know, I think one of the issues that a lot of marketers have with AEO is that there's not a lot of predictability in the results that you're going to get. Right. So if somebody goes into, you know, ChatGPT or Claude and asks for some comparison data, you could get three or four products listed and then you do the same search again. And there are, you know, different products that mix the next time. If you or I were to search for the same term, obviously we would get different results because all of the variability in our data, right. That that AI is collecting. So trying to aim at that target feels impossible. How do you think platforms like Octolens and sort of being more broadly aware of just like the conversation, the community, you know, how have you seen that work for folks and sort of what are they doing to sort of combat that sort of variability gap that a lot of people are running into?

[30:18] Charlotte Schmitt: Well, one aspect of that variability I think is data freshness. So AI search will index quite high on very new information out there. So then that would be another reason to say, okay, you need to be on top of the new trending threads, the things that are important today. So that's one way of at least staying into in this data freshness bucket. And then overall, social listening will help you understand how your users are talking about your, your product, your industry in their terms and their language, which I think will also help in understanding actually what prompts you're optimizing for what messaging and what keywords. So you can also draw from what real users are saying out there to help understand really what you're optimizing for an AI search. So those two things can help for sure as well. And then outside of the AI search aspect of it, I'm still being super active on the communities and being present there might give you more stability than counting on ChatGPT exclusively as a platform to help with distribution.

[31:31] John Steele: So let's talk about AI more broadly as a founder and using that within your product. So obviously we talked about the variability of sources, but also like different models that come out, right? Different models change from day to day, from week to week. They can have these drastic sort of swings in sort of what they're good at, what they're not good at anymore. Does it make it difficult as an engineer building something that is, is sort of using AI, you know, relying on AI to a certain extent to sort of help with different processes? Is it hard to rely on AI as sort of a backbone for the underlying engineering and sort of how do you see that going as we go into the future here?

[32:16] Charlotte Schmitt: So in Octolens, one of the main ways that we use AI is telling a user what posts are relevant and important versus others that are not. So if you do keyword based social listening, one of the things is or the one of the hardest parts is just surfacing what is actually important because there's so much junk. There's so many things you don't want to actually be paying attention to. So this, we call it a relevance filter is one of the main ways that we use it in our product. And for us, honestly, just over time, it has gotten stronger and stronger and better as new models come out. So that's been great for our product. We did recently, for example, there was this big hype for [unclear: model name] and [unclear: model name] is was or is a really great model for classifying data. So we thought that would be amazing for Octolens because they can help with our yes, no relevance. Turns out not so much in our case so far. So we have to maybe spend some more time optimizing it. But for our product so far, it's been, you know, just getting much better over time.

[33:18] And we have actually been able to stay just two full time people at Octolens. So incorporating AI in our product, but then also our internal processes has allowed us to stay just super lean. And now it's been two and a half close to three years. And in the beginning, I would have thought by now we would have been a much larger team at this size, given how many customers we already serve. But I just like keep pushing back, thinking about hiring anyone because we can do so much just as a team of two.

[33:48] John Steele: Yeah, I do want to I do want to come back to that because I do find it. I was going to mention that earlier that it is great that you guys are able to still be a team of two. And I do want to come back to that and talk a little bit about some of the ways you're doing that. But I've been struggling with something a little bit where, you know, with AI and people being able to sort of spin up different products or tools for themselves. Do you struggle at all with what parts of your product are sort of your own and cannot be duplicated and sort of what you can protect, what you can sort of own, what your what is your intellectual property when, you know, somebody can go into a platform and just sort of be like, hey, spin this up for me. You know, how do you guys sort of see that? How are you able to protect against some of that?

[34:39] Charlotte Schmitt: Yeah, well, I think a lot of the protecting on our end is that we combine so many different data sources into one. And even though, yes, it's getting easier and easier to vibe code, especially a dashboard, you know, or your kind of your your SaaS tool, it's still very difficult to handle, for example, data pipelines getting like really clean, relevant data to the user at the end. So we have so many sources that we monitor and only serve you the important ones for your business. And that's taken, yes, the three years, but then also the years before that, when I was a crowd.dev of work on all these different, yeah, data pipelines for the product. So that's one big aspect. We've had several customers or users who have come to us and say like, yeah, we've, we tried, we tried to build this on our own, but in the end of the day, like we, it wasn't worth our time anymore to try to maintain. All of these different data sources and get actually good clean results at the end. So that's one aspect.

[35:40] So it's actually not that trivial to, to copy. So being this clean data infrastructure layer is what a lot of our users value. And like I said before, a lot of our users are very technical. So if this was easy to rebuild, I believe they would already now. So that's, that's nice to have that, those many years of experience in setting up these different data pipelines. Yeah. So that's one aspect and on top of that, just being ready for the next wave. So having an MCP, making sure users can plug into their tools super easy, having this read and write API, that's what we're focusing on. Really staying on top of what do users want now from, from Octolens and staying on top of, yeah, serving them those features is what we're, we're focused on to stay relevant. Yeah. And actually it was so funny because I think it was a couple of weeks ago, someone came to us and wanted to do or get our data into their platform for having social listening in their marketing suite.

[36:41] And we said, yeah, sure. Let like pull up also your product so we can see where the data ends. And they had literally just copied the Octolens dashboard. And I think they did this probably like, I don't know, a year or so ago, because it was one of our older versions. And then they were asking us for the data aspect. So there you can see, okay, you can vibe code a dashboard, but then actually getting the results or the data in is not that easy. So they didn't even, the person on the call hadn't realized that that was our UX that they had. Um, funny moment that they couldn't actually make the data aspect work. So yeah, not that easy.

[37:16] John Steele: So last week on the show, uh, I had a marketing person on who had worked with a company that focused on developers as well, building on the platform. And they were building a community of developers. And then when they sort of went to expand and they sort of wanted to drive a more commercial motion, they went to broaden their appeal. And I had asked her, you know, why do, why does she think that is? Um, and I wanted to get your read on that too. So do you see yourself sticking with like your, I'm not, I know it's not your complete customer basis, all developers, but you, you clearly started from there and you clearly have some partnerships with some developer focused. Data sources. So I'm curious as you sort of look to expand the company, uh, do you see yourself sort of broadening to different groups or do you see yourself sort of staying here and sort of figuring out ways to go to enterprise, go up the chain, go to like the CTOs or something like that?

[38:10] Charlotte Schmitt: Well, yeah, we've definitely gone beyond dev tools and it just happened kind of naturally because I mean, one of our main growth drivers was just luckily word of mouth. Um, and I think we were also, you know, just being talked about in like meetups and marketing meetups and go to market engineering meetups and that just increased our visibility across different segments. So it's really tech brands right now that we serve going up the, like to corporate. I think that's difficult as a team of two to be a completely like sales led company. It is as a small team, a lot more doable to do self serve. And that kind of motion, if we decide now to go like full corporate and start going into these six months procurement processes, I don't know how long we could stay a team of two and it would be a different company set up. So that's not exactly what we're focused on, but I mean, definitely the aspect of focusing beyond dev tools has come naturally after I'd say the first year. And so now I would say it's tech teams and that includes, you know, broader B2B SaaS and AI companies that use Octolens.

[39:20] And it's just the companies where users talk about them and opinions form on socials is our target.

[39:26] John Steele: Yeah, we're definitely seeing a broadening as well of, you know, who would be deemed like technical, right? Because I think like a lot of people, you know, in a marketing or sales capacity even or a more operations capacity, you have to be very technical if you're working at like a SaaS platform or something similar. So, yeah, I can see that how, you know, you're sort of marketing to the same folks. But do you think that to the extent that your platform works with these sort of things like Reddit or Hacker News, for example, the audience is there would need to diversify. Yeah. So how do you sort of see that happening? Do you see yourself like expanding the data sources that you guys are listening into? Or do you think that Reddit becomes more popular with more types of people, for example?

[40:13] Charlotte Schmitt: Well, I think Reddit already is popular or it's one of the biggest social or community sites beyond.

[40:20] John Steele: Yeah, it's maybe a bad example. Maybe, maybe the difference between X and Instagram might be more of a, more of what I'm going for here.

[40:27] Charlotte Schmitt: I think, but you're, you're definitely right. And last week we actually added Instagram as a platform. Oh, really? So, yeah, so we have, we also already had TikTok for a while. Facebook is coming next. And that is because our customers getting more broad. It's not just the ones that care about X and Reddit or LinkedIn, B2B SaaS. But also because B2B SaaS does now care about places like TikTok and Instagram more because reels or shorts are becoming important also for B2B SaaS companies. So, and it's, it's widening because our existing customers care also more about these channels. But then yes, we have just more companies coming in asking for these more traditionally B2C sources. Agencies also have been becoming a bigger group of customers and they'll have different customers with different needs, including places like Instagram. So we've added that for platform coverage and, you know, podcasts, important for everyone, news, web. So at this point we cover quite a large chunk of the web.

[41:32] So it goes well beyond DevTools.

[41:37] John Steele: So you're a team of two, you're serving hundreds of customers. Talk a little bit about the tools you use to keep this all going. What's sort of automated? What do you guys sort of still do yourselves? You know, and how do you, how do you sort of like balance the division between the two of you and sort of keep it going with, with such a large customer base? Yeah.

[41:56] Charlotte Schmitt: So I am basically in charge of everything non-engineering and technical and Igor, our founding engineer, he is then doing technical side of things. And I mean, it's also great because, you know, a couple of years ago, I wouldn't have been able to take care of all, you know, just all the website changes myself or that all like falls into my bucket now as well, because I can just do that via Claude. So, so yeah, I have a lot of my processes automated at this point in Claude and we did that almost two years ago, I think. And we keep going with this. We always sit down and write down all of our different workflows that we own and see how much time that's taking up in the day and how much of that can be automated. So there's a lot of things nowadays which used to take several days, something like let's say accounting. You might have even need to hire a person for that, but now that's so much automated and, you know, also content tasks have gotten so much faster because you can get first drafts done for you.

[42:59] Our customer support has become much more manageable because you can get such good answers for customers already, SEO topics, everything has just been, you know, you can get done so much faster. So I do most of my work with Claude and I like Claude Cowork and there I have my, I don't know, 12, 15 recurring or scheduled tasks that I then just like power through during the day. The results that it gets so that I can do the final edits, for example, in content and that's been able to keep the team super lean and of course, Igor on the developing or development end has gotten a lot faster as well.

[43:37] John Steele: Hmm. You know, we talked a little bit earlier about what, what intellectual property looks like and how you create a name for yourself, differentiate yourself when AI is so injected into everything. What's the line for you in terms of process? Where, where do you see, okay, I'm not going there with AI. I'm not letting AI do that for me. Um, you know, and how does that work when, you know, when maybe you and Igor have different sort of like opinions on that? Like, do you, do you think that there's, there's a line for you and your company as to, you know, where AI is going to go for you and where you're going to keep you the human sort of in the loop on it?

[44:15] Charlotte Schmitt: Well, I would say human in the loop is almost always there. I'm not someone who will just like get your AI generated LinkedIn posts and just like just posted without even reading it first or our blogs or I think you do need to keep that quality bar high and have that human in the loop. Aspect. And I always try to do everything like until I know it's working perfectly well myself before even trying to automate it and hand it off. And for example, just last month I was, we were looking at our activation and from free to paid how that could be doing better. And we also wanted to rebuild part of our onboarding. So to have that done, that was me just essentially the month of September, going through every single signup, qualifying it, going through their workspace, fixing it, sending them a manual outreach to try to get them on a demo call, have a demo call, like super manual tasks, because I needed to 100% understand, okay, what's still missing in this free trial week that we need to perfect.

[45:20] And now in the first week of October, I've been able to hand 90% of that flow off to Claude because now I understand, okay, these are the five things that are the problems and need to be fixed. And then I only need to in the end, you know, do the tweaks of the email before I send that to the customer to get them on a demo call. So it's about perfecting it and understanding the flow yourself first before even trying to automate it. I think that's one big aspect. And this, this human contact for a startup or a company of any size, I think is still so important. Like if I can spend my whole day on demo calls or calls with customers, that's an amazing day spent. And like, and that's not something that I think will be or can be handed off anytime soon. And especially as an early stage company, like you need to have that direct contact with your customer and not try to automate away all of customer support, all of that communication. That's something that's just so valuable. And I, I own that for, yeah, for myself.

[46:20] John Steele: So working with developers is an interesting place to be in terms of, in terms of that, right? In terms of AI, because you would imagine that developers are using AI so intimately throughout the day. They're trying to inject it into all these different processes. You know, some folks even talk about, you know, AI, you know, in a more global way, more of a general intelligence or something like that. But then on the flip side of that, you had mentioned earlier, and I know this to be true, developers being allergic to marketing speak. And I imagine that extends to the way AI speaks as well. So how do you see that sort of dichotomy? We have AI working within our platforms. You have AI delivering certain things, but, you know, how do you balance that with, with such strong opinions and, and, you know, sort of aversions to it in some cases? Yeah.

[47:08] Charlotte Schmitt: So I actually also just wrote about this on my LinkedIn yesterday or the day before. So there's a whole segment of tools, which kind of stem from social listening tools that promise you like the customers on autopilot from Reddit, let's say. And these tools will take it a step further. So they'll do the listening part, grab posts, but then they'll also have essentially bots created or user accounts on Reddit. And we'll then also do the final posting and say, you don't even need to touch this whole flow and we'll just automatically get you customers. And we very purposely, and also because I think it just won't work long term, I've stayed away from that completely autonomous workflow. Because in the end of the day, especially on these community platforms, like humans want to talk to other humans and want to get recommendations from peers, want to read thoughtful and authentic content. And the companies doing this right and using social listening as a tool to find the right threads, but then doing that extra or that last step of putting in value in their own voice or the ones that are really doing well and winning.

[48:13] So there is that line of like, yes, try to automate the mundane, the, you know, you don't want to go to every social platform and search your brand name and find the posts like that. And those routine tasks should be automated. And that's what we help with Octolens. But we are not the tool that says, okay, and now we'll give you the answer and post it for you because I think first of all, it's not going to work long term. And yeah, that's just not our, that's not our approach or what I think will make a brand strong. It'll just tend to, you'll lose credibility and it'll just be harmful long term. So I think we have a strong line at that point.

[48:50] John Steele: Well, that actually makes sense. It's social listening, not social talking, right? So you're trying to, you're trying to hear what people are saying, not necessarily speak back to them. And you leave that up to your, to your customers. I think that makes a lot of sense. So it's the question that we ask every guest that comes on the show. Our final question, what is a integration automation workflow or tool that you can't live without?

[49:13] Charlotte Schmitt: Well, the one that I have right now, which I can't live without is, excuse me, the one I just mentioned around the understanding our trial users. So every single morning in Claude, I have a recurring task where it gives me the rundown of who were our signups from the last 24 hours. What were the actions they took in our product in the beginning? And that's via the PostHog MCP, so our product analytics tool. And then on, on top of that, it'll suggest me an outreach to try to activate that user. And then I'll go in and just do that final step of, of polishing that email. So I'm combining our product analytics tool with our own customer or with our own tool to give me that information on our new users. And that's what I can't live without right now, because it tells me, you know, like what's working, what's not, and it'll turn out for them to have a much higher conversion rate if I go in there and address each user personally.

[50:15] John Steele: Well, Charlotte Schmitt, thank you so much for joining me today. I really appreciate it.

[50:19] Charlotte Schmitt: No worries. It was a lot of fun. Thanks for having me, John.

[50:21] John Steele: All right. Thanks a lot. Have a nice day.

[50:23] Charlotte Schmitt: You as well. Bye bye.