← The SaaS Podcast - Growing Profitable AI SaaS & AI Agents
The SaaS Podcast - Growing Profitable AI SaaS & AI Agents
The Risky AI SaaS Rebuild That Broke a $2M ARR Ceiling
00:55:02
Karel Papik (Product Fruits) on rebuilding an entire SaaS platform around AI, the gaming psychology behind 24% free trial conversion, and why PLG stopped working at $2M ARR
Loading summary
Transcript79 lines
- [00:03]
A
Welcome to the SaaS podcast. I'm your host, Omer Khan. AI has changed the playbook for building and growing SaaS. Every week I talk to founders who are writing the new one. My guest today is Carol Papik, who built Product fruits to over 1300 customers while competing against companies that had raised hundreds of millions. His total funding, three and a half million dollars. But when AI competitors turned up, Carol panicked. Two weeks of what he calls dark times. Then he emailed his investors and told them we're going to stop working on the product and we're going to rebuild it around AI from scratch. Expect revenue to decline. Meanwhile, their product led growth had already stalled and their PPC had hit a ceiling. In this interview, Carol breaks down how a marketer spending just one hour a day turned PPC into one of their biggest growth engines. And how he's solving the AI pricing problem when his costs are usage based but his customers won't pay anything but a flat monthly fee. I talked to a lot of founders stuck in the same spot. They've got a clear vision. They just need the right team to build or scale it. That's where Gearhart comes in. They're an AI powered product development studio that handles the entire technical side of building your B2B SaaS platform or AI agent. Built by serial entrepreneurs, they understand the unique challenges of startups and can plug into your team to accelerate growth. They've built over 70 successful products, including SmartSuite, which raised $38 million and is used by companies like Capital One. Right now they're offering our listeners the first 20 hours of development for free. Just book a call at Gearhart IO. That's Gearheart. If you do SEO for your SaaS, you already know AI is changing how people find products. Buyers are asking ChatGPT and relying on Google AI overviews for recommendations. Instead of browsing traditional Google search results, Respawna helps you show up in those AI answers. They get your brand featured on the sites that AI tools actually cite. They helped Opus Clip hit number one in AI visibility and add over 100,000 visitors to to their monthly organic traffic. SaaS podcast listeners get $1,000 in credit when you sign up for a subscribe and save plan. Just mention my name after you sign up. Visit respona.com to get started. That's R E S P O-N-A.com right now, there's a number hiding in your SaaS metrics. The exact point where your growth will stall. You can change it, but first you have to see it, it's simple math. When, when your churn catches up to your new revenue, you stop growing. You're running just to stand still. I've built a free calculator that shows you exactly where your ceiling is and the fastest way to break through it. Find your ceiling at SasClub IO calculator. That's SasClub IO calculator. All right, Carol, welcome to the show.
- [03:02]
B
Hi. Thank you for inviting me, Omar.
- [03:05]
A
My pleasure. So are your target customers SaaS, AI startups? Manybody.
- [03:13]
B
I think we have more than 1,300 paying customers now and a lot of them are SaaS companies. Also a lot of startups, but we have universities, insurance companies, companies which are building ships, TV stations, stock exchange. So it's really much wider area than I actually expected at the start. At the start, I expected the use case will be improving conversion in free tr, but I think we actually have now minority of customers which need this use case.
- [03:44]
A
Great. And give us a sense of the size of the business. Where are you in terms of revenue, size of team? You already mentioned customers.
- [03:51]
B
Yeah, so yeah, 25 people, mostly in Prague, Czech Republic and also in Pilsen, Czech Republic. And 1,300 customers. Biggest market US, we are in regards of ARR, in lower millions, I would say, and growing quite nice. I think, at least in this particular moment, we are the most innovative AI platform in the world. DAP platform in the world.
- [04:26]
A
Okay, great, we'll get into that. And you're wearing the product Fruits T shirt. What does it say below that?
- [04:32]
B
Riding the tiger.
- [04:34]
A
What does that mean?
- [04:35]
B
Well, it's a saying in Asia. I spent some time in Asia. I used to live in Vietnam several years and other countries and well, they have some saying that in some moment you can ride a horse. Right. And horse is slow and sometimes it's actually dying and you don't see it. And I think it was kind of product for two, three years ago. We were riding the horse. Everything seems like peachy. It's fine, you know, competitors here and there, everybody fine. And then you are thinking, wow, there are some new guys behind us on motorbikes, AI powered. They will take everything from us. We have to regroup and start to think very differently. And that's what we did. And now we are not riding horse, we are riding a tiger. And when you are riding the tiger, the tiger will eat and destroy anything and everything in front of you. But you can't stop. If you stop the, the tiger will eat you. So the only way is to aggressively continue.
- [05:33]
A
I love that. Love that. All right, so let's get into the story. Like, where did the idea for product Fruits came from? You founded the business in 2020. What were you doing at the time? Where did the idea come from?
- [05:49]
B
Well, my background is in video games, right? I did video games all my life. So I also learned a lot about onboarding and adoption in games, which are still, I would say, light years ahead of B2B applications. And when I returned from. From Asia, I acted like. Act like some advisors for some startups. And one day I. I met Ladya, and Ladya is the brain of the company. I'm just talking head, right? I have no idea about anything. But he is really brainiac. And I, you know, honestly, I don't respect people much. That's my personal problem, you know, But I respect so much Lagia. It's very uncommon for me. You know, sometimes you meet somebody and you. You see and feel he's special. And he was alone doing product fruits, and he had, I think, six customers or something. And I was talking to him like, 10 minutes, and I. I almost said, will you marry me? Like, okay, let's do this together, right? So we started to do it together about one year. We tried to bootstrap it at the start, and then we got some investment and the ball started rolling. So the idea is, Laja's idea is to shy, to go to the. To the podcast. Me too, but he's even more shy than I am.
- [07:22]
A
So I'm curious, when you had that moment where you were like, I'm in. I want to do this. What was it? Was it him as a potential founder you could work with? Was it the idea of the product?
- [07:38]
B
I strongly believe in the product because I'm using all kind of software every day. Vast area of software. I'm doing also some. Some art. So I'm using, you know, animation software and graphic software, you name it. And I. I always struggle how to use it right? And I always needed some help. And I said, this is it, this is it. This is. This is what I always wanted. And he's so calm and he's. He's gifted because he always, you know, a lot of founders, because we are, you know, grinding every day, some here and there, it's just fighting fires, but he always is able to see the future, like one year from now, two years from now. And he's always right. Big respect. So this is it. He's winner.
- [08:28]
A
So I'm curious, with your background in gaming, you know, I work with somebody on my team when I was at Microsoft, who had spent as much time as you have, you know, 15 years or something working in Xbox. And he brought a completely different mindset to building B2B products or web products. And so I'm curious when you, with your background and all the things you talked about, when you looked at product fruits or you just generally started looking at other SaaS products, what did you think?
- [09:11]
B
They suck in adoption, in onboarding. They are hopeless. They are rigid, they are too shy to talk to the customers and they are sometimes even afraid to try to feel what customer is feeling. And this is something what you learn in games. I shipped I think 16 games and a lot of them are actually for women. Surprise. I had a chance to work on TV franchise. I don't know American Castle, if you know or Criminal Mind. So I made games based on this. And we always had the first hour for free. So in first hour you had to convince the player that she want to buy the game, right in the first hour. So you had to explain everything enough during the first hour. But not like, not to be too pushy. So it was big learning. I will give you example like in SaaS. Okay, so you have some, some tool set and you are trying to upsell the customer. So customer is hanging around. Maybe he will upsell because he needs, I don't know, automatic translation, whatever. In games, how it works. You are playing game and you are, I don't know, killing zombies, right? And in one moment I will give you Diamond X and you have the ax and you are killing zombies like crazy. And it's so much fun and you enjoy. And after 10 minutes I will say, okay, Omar, give me back the ax. You can't it unless you will pay me. If you will pay me, you can keep it. It was great. It was great fun, right? You enjoy it, but now you have to or you pay. So it's a different approach than when you will ask for the money up front and some dedication. So some, some game principles are even forbidden, like complete Gotcha. It's a gaming principle which is forbidden in Japan because it's. It works so great that people are pouring money to the game. So yeah, there are some techniques which you might want to implement in games.
- [11:12]
A
So I know people talk about gamification is sort of one thing, but what you're talking about there is a lot of psychology.
- [11:20]
B
Yeah, it's a lot of psychology because game is asset which you don't need. You don't need it. So the onboarding must be really very well thinking through it, because it's about emotions and stuff. So this is it. In a lot of B2B software, you just need it, I need it. I need something. I can choose this, this, this, right? But in games.
- [11:48]
A
Give me an example of the axe. Example you gave with the game. Give me an example of how that translates to something that you're doing in product fruits.
- [11:57]
B
Well, one of the features is a copilot, right? I would like to avoid to say chatbot, because our chatbot is much more than a regular chatbot, but he's answering some questions, right? But he's also able to point on the screen, tell you you want to try this, try that. You can talk to him via voice and his memory and stuff. So it's really, it's really great. And it's. But it's kind of expensive. It's not that cheap because it really delivers. And this is what you need. You need top quality in answers, nothing less. And we gave our customers just few, I would say tokens, or we call that resolutions, you know, to, to try. And it was a barrier. And then I said, okay, so fuck it, you know, we will, we will give them zillions, zillions. Do anything you want to. We will cover the expenses. Because you want to try. Because we know that if they will try, they will become customers because the value is there.
- [13:03]
A
Sounds great. Also sounds like a recipe for running out of money really quickly.
- [13:08]
B
It's risky. And we had some calls also, for example, with Gartner, and they said, this is the way you should do. That's the only way. Because, well, as I told you, we are, and that's a fact, the most innovative AI platform in the world, right? And we are selling quite well. But there are still a lot of barriers. A lot of barriers. We are a rather small company, right? And when we announced some new features four months ago, we had webinars and I was attending 400 companies, which is quite a lot for us, you know, and the big companies, KPMG and stuff, and everybody interested. But still there are some barriers for better adoption of AI. And it's mostly about the human mindset. I think the word we are looking for is uncertainty. They expect computer software to be very deterministic. One plus one is two always. That's it. But this AI is not like it's stochastic. You know, there are probabilities, it's a bit random. So some of our customers are saying, oh, we are calling him Alvin when he's on a chat. He can do many things. Discovery calls Outcomes, many agent identities, I would say. But when he's responding on the, on the chat, he's giving you some answers. But the answer doesn't have to be always the same based on the context and stuff. And some of our customers are saying, but I want him to, to say exactly this. Theoretically we can't, we can do this for particular client, but we never did because that's not the way to do it. If you want to scale, that's what I'm saying to our customers, you have to accept the uncertainty. You will never have it fully under control. So of course we have some processes like I can go through the, or our clients, let's say I, we are selling product fruits to some client and client uses our chatbot and he can go through the conversation, say this is good answer. This is, this is wrong. Next time you should say something different. You have to tune it up. Same for pricing because pricing is very tricky with AI. We are paying shitloads of money to different providers which orchestrates all this thing. And when we are selling the company to some decline in there. Okay, so how much will we pay? And our answer used to be we have no idea. We have no idea because I don't know, how many users do you have? How many of them will, will interact this product fruits? How many, how many questions will they have every day? Two questions, 20 questions? I don't know. Right. So I will, we will give you some. We will, the original idea was that we will charge you based on some resolutions. Right? Then we will provide the right answer. And it didn't work because you know these people always said, but I need some number. I will take this number. I will go to procurement and they will approve the purchase. So they want static price setup price monthly or yearly while we are paying per conversation to the LLM models and else.
- [16:42]
A
Yeah, I know that's a big issue right now that we're hearing that when it comes to SaaS per seat pricing is, is dead. It's a problem and for many it is. Although I still believe that many customers don't care what the pricing model is as long as they feel like they're getting value for what they're paying for and some level of predictability, which is the other issue. If you go completely towards usage based pricing and say, hey, this month it could be $5, next month it could be $500,000. Right? It's like, I don't think anybody wants that level of uncertainty as, as well. Let's talk a little bit about how the Days when you were riding the horse, not the tiger, right? So when you joined there were like six customers. Walk me through how you got to the first 20 or 30 customers. Where did they come from? How did you acquire them?
- [17:44]
B
Well, I often, often read that there is certain paths from 0 to 20, 20 to 100 and this is something what I am trying to avoid. I think it's a bit wrong. Let me explain what do I mean? Like especially we are from Europe, right? So Czech Republic. So a lot of, lot of startup here. They are starting on Czech market. They will ask for some precede money and then they are trying to expand to Slovakia, Poland, step by step. And it's. That's usually failing, right? It's failing. You want to know in my opinion from the day one that you are able to succeed on the big market. The big market is us. If you are starting in US very early, you see that your product is able to compete with the best companies in the world. And it's not just about the product but about the whole sales engine, marketing engine. I met some founders which were super great selling personally and even huge tickets, right. But then they couldn't expand because they were not able to teach their sales department how to, how to really sell their enterprise software because it was so complicated. So from the day one we were looking for some, some path. You know, we, we asked for the first investment from Lighthouse Ventures. It's just great Czech Farm. And, and they told us okay, so you have certain time, maybe year to, to to show some numbers. You need to, to grow like it's from zero. But you really need aggressively grow. And you have 12 months to see the result. And then you, you can ask for more money on the market. If 12 months, how will you do it? That was the times like three, four years ago. Everybody was obsessed about content. Content, content, CEO content. And I would like to have content agents, you know, these people that okay, we will help you with the content. But it will work after nine months it never did. So they will go to another, another. So we said okay, so no, no content. We will Simply go with PPCs small team and this will be the. And from the first day we will try to sell in US or at least UK and Germany and so on. Still today we have just, I don't know, 10, 10 companies in Czech Republic. Sales department doesn't speak Czech at all, right? No, they don't. So that was the past. We tried to scale from 0 to 10 to 50 to 100 in the same way. It should be the Same momentum. And this worked I think until 2,000,000 ARR. 2,000,000 ARR is common friction point when a lot of us starts to struggle because they are unable to break into the mid market.
- [21:04]
A
So before we get to that, I want to just talk a little bit about that 12 month period. Where were you at the start on day one and where did you hope to be at the end of the 12 months?
- [21:21]
B
At the start we had this, I think seven, 10 customers and within the. So the MRR was I don't know how much, next to next to zero. And I think within the year we had like 50,000 mrr or something.
- [21:41]
A
So by the end of the year you were doing about 50k in MRR on day one. When you started this 12 month journey, where did you hope you would be
- [21:52]
B
approximately at this number?
- [21:54]
A
Okay, so walk me through how you got there. You mentioned ppc, was that the number one thing that.
- [22:04]
B
Yeah, because we couldn't afford a big team. Right. It was very, very small team, five people, four people. And we couldn't build a team which would create the content and more sophisticated lead channels. Actually we had like first hire was Dan, then Musial, he's great guy and we hired him as a chief marketing officer which was lie. Right? It was complete lie. He's not marketing officer, he's plumber. He's plumber. He's fixing everything here and there without even telling us, oh, that was a leak. So I fix it. That was a leak, fix it. And he could work on the marketing like 20% of his time. So like one hour per day max. So that was his marketing. So also that was the reason why we started with PPCs because for PPC you just need money with some level of simplicity. Right. He was hired like a marketing officer but he could afford or we could let him work on the marketing only like one hour per day because he had so many other stuff to fix in the company, you know, many, many others because we're very small teams. So he said okay then you can, you can only focus on marketing, I don't know, one hour per day. So PPC is the way to go because you don't need that much manpower because you have so many other stuff to do.
- [23:36]
A
Now when it comes to PPC, was this AdWords or something else?
- [23:40]
B
Yeah, mostly AdWords. And that's another like I would say problem because at the start it worked quite well. We had conversion in free trial, I think 25%, 24% which is quite good. And based on the calculation and size of the ticket at that time, the ticket cost, the average ticket was $100. So with the conversion 24% to the free trial, we had our money back in nine months. Eight, nine months, which is good. Right. And so we scale the PPCs until I think $1 million yearly, a bit more, maybe 1.5. And that's a ceiling. You can't have more. It has limits because only certain people are searching for the solutions. So you have to. But at the start it worked great.
- [24:35]
A
So 99% of founders I'll talk to will tell me. We tried paid ads, PPC and it didn't work. So we did something else.
- [24:44]
B
I can tell you why it worked.
- [24:46]
A
Yeah, I want to know that. And I think there's one other thing that I wonder about, that not only getting it to work, you're getting it, you're trying to do it in the US market, which probably the clicks are going to be more expensive than if you were trying to do it in the Czech Republic. And also you've got some very well funded competitors who could probably. Who charge a lot more than you do. We'll talk about pricing in a while so they could afford to spend a lot more on acquiring trials. So with all of those factors going on, why do you think you were able to make this work?
- [25:21]
B
So one of the main reason was Dan, because Dan was really expert in PPCs. But when we hired him, he was the PPC guy, right? So he knew how to orchestrate it. And we still do. We are running lots of landing pages and stuff. So he knew this game, he was very good in this and he was able to place us among the customers. And actually that's what I was mentioning before that we wanted to know as soon as possible that we are able to compete. And to be able to compete it does mean that you have to have the good product, but you also have to have some marketing which works. We wanted to know as soon as possible, like if we are able to pay for the PPCs in some meaningful way. Right. If it would be too expensive. So we didn't want to start on Czech market in Slovakia and then try to go to the world and say, oh, it's so expensive, we can't afford it, we can't afford it. We wanted to fail as soon as possible. We wanted to see if we are able to survive and if not, we wanted to fail as soon as possible.
- [26:34]
A
So do you think it would still work today? I mean, every year I hear people say, oh, PPC bidding for this stuff is like so Expensive now, but I'm sure in back in 20, 20, 21, whatever it was, people were saying the same thing. Oh, PPC is too expensive, Clicks are too expensive.
- [26:59]
B
Right. You know, we are also using some software which is tracking how we are doing in when people are looking for some solution using, using AI, you know, in AI browsers or in GPT. So we are trying to tune our content now to reflect that. But what's cheaper than PPC's content? It is not. You need to have bunch of people working on the content or direct sales. So you need to have bdrs sdrs. So nothing is really cheap.
- [27:31]
A
Yeah, I saw an interesting tweet today. Somebody was talking about this AI tool that could basically help you create thousands of content articles in seconds. And somebody kind of responded to that and said okay, great, but now when everybody can generate thousands and thousands of articles or content, why are they going to pick yours? You still have the same problem. You still got to figure out how you stand out. Right. And so I think that that's the thing about content. It does an SEO and inbound, it takes time to get that to work. And now I think it's probably even more difficult because there's so much AI slop and so much noise out there.
- [28:30]
B
True? Absolutely true.
- [28:31]
A
Okay, great. So let's talk a little bit about pricing. You charge from. I think someone can start using product fruits for about $100 a month and then you have higher tier plans but you've got competitors, competitors who charge a lot more than that. Tell me about like why you price the product the way you did and are you ever in a situation where people complain that the product is too cheap?
- [29:09]
B
Yes, exactly. We talked with some recently to some Israeli company and they said that there must be some. There is something what you are not telling us. It's. This is not possible. We tried our software, it's amazing. Why it's why it's so cheap. It's significantly cheaper than the Pendo. What's the problem? There is some hook. You know me historically we started like a company for small SMBs companies and the price was very important. Right. And then actually we saw some. I told you that at the start we had conversion like 24, 25% and we used a lot of PLG principles in the application because we are digital adoption platform. So we try to benefit from having product fluids in our product fruits, if you know what I mean. It made a lot of sense and it worked. But we saw steady small decline in the conversion. You know, the leads Started to converting worse and worse all the time. Now we are around I think 15% which is still not bad. And the reason for that is that the platform, our platform fundamentally changed. At the start it was quite simple platform you can operate now. Today it's so complex. I just, you know, five years ago we had the product tours and hints feedback maybe, you know. Now our AI is able to have discovery call with you and ask you what you are looking for. And then based on this it will tailor the onboarding. You can ask AI like what I did today, 10 o' clock in the morning I wrote to our AI, our administration which we are using on in our product which is kind of confusing. So what is preventing us to see better conversion in newsfeed feature. And our AI went through all this conversation and told me this is the problem and this is how you should fix it. And this is also what we are selling, right? If. If you do the survey with our competitors they also sometimes they are offering survey. You have to set up the survey question. This is with us. You will just ask Elvin, hey Elvin, please put together some churn survey. And by the way, last time I did that and I said okay, please change the color to red. And he said Karel, I don't recommend it, it's a scary color. I said okay, so let's go with green. So yeah, so now a lot of things in the product fruit is very complicated and PLG stopped working Practically you are not able to sell through the PLG. So we had to start to build some sales department still quite tiny 15 persons including implementation engineers. And we grow not because we are acquiring more and more companies, but the tickets are bigger. So that's what we want to change. When you are hitting around 2 million ARR, you can grow by still adding more and more customers as you used to because you would need too much people in CSM and tech support and stuff. So you need to start to sell the bigger accounts and for that you need to have more complex, more complex enterprise ready product. So that was the change in our, in our journey. It worked quite well powered by PLG and PPCs until I would say 1.7 million, 2 million ARR or something. And then it started, you know, to decline and. But we had that scary discussion with Slaadja about a year ago. I remember what we will do for the future.
- [33:06]
A
So okay, so one thing I just want to understand is why do you feel PLG stopped working? Was it just because the complexity of the product?
- [33:16]
B
Exactly. PLG works for. For simple products or work the best for the simple products. But if you have more complex product, it's not it. It works only to some extent. Because for example, when we have a call with a client, it can be big client and they have some woke idea what they would like to see in a product adoption. Because we are, we are doing much more than onboarding. You know, we are taking care about the customer in the whole life cycle. You can publish announcements inviting people to the webinars and people are sending you videos when they have a problems, churn survey, you name it. So sales guys are account executive. They need to sit with the customer and put together the use case. And almost always the customer is like wow, what is even possible to do now? It's amazing. So they don't know that they can't expect it. So PLG doesn't work that well.
- [34:16]
A
Got it. Okay, now at the beginning we talked about, you explained about riding the tiger and what that means to you and the business. And that before you were riding the horse until you realized these motorcyclists were speeding and chasing you. These AI guys. Tell me about that moment where you guys realized that they were behind you and that this was a problem.
- [34:42]
B
We were scared. We were really, really scared. Like depressed and disappointed. Like I said, we did what we could and we did right? And now world changed and changed so fast. And there is AI and we are not AI platform. And it was dark, dark times. But it didn't take too long. It was like two weeks. But I remember every single day on this. And then we said, you know what Laja? Let's think about the platform which we always wanted to build and we couldn't because of some technical barriers. Maybe these barriers are not here anymore. So let's start dreaming, just dream what would be the best adoption platform not just to have AI because investors want it, but something. What is really amazing what will help people working with the software. Because the end game in my opinion is not having some adoption platform which help you on board to some platform. I will give you example Photoshop, you know this graphical program for graphic, right? So you can have some onboarding to this platform. And it will say, okay, you have this tool, that tool, this, this, this. It's so, so many things. But what we, what we are building now is the invisible body which is sitting next to you. And when you will be working with Photoshop or anything, you will ask him, hey, how to. How to paint the sky? How should I paint the clouds? And he will tell you, oh, for clouds you can do, you can use this tool, that tool, this tool. Or maybe I can try paying the clothes for you. So the buddy is always here, always helpful. Right? And this is what we wanted. And I remember I wrote some email to our investors, which are great. You know, we have Lighthouse vf, Reflex Capital leverage and I, I send them, guys, we are stopping work on product fruits. We will mostly rebuild the platform from the scratch. So I expect you will see decline in MRR because we will not update the tools we have anymore because this is not the future. We could grow in next six months, nine months, maybe year, but then we will be dead. And we see this point. So we will now pour all our resources to the new generation of the platform. And I hit the sand and said, okay, so now let's wait for the shitstorm. And they called me in 20 minutes. I remember Onze from Reflex Capital and he told me, how much money do you need? We are betting on you. This is what we want. We want winners. We don't want survivors. Tell me how much you need. And was great, you know, so this is what we did. We actually didn't see any decline because the products still work very well. But from September last year, July, we started to introduce the new level of tools which are AI powered, not just AI spies here and there. And it's amazing. It's almost scary.
- [38:15]
A
Yeah. And I think once you made that pivot to everything being AI first, you saw some significant growth as a result of that. But I'm curious because these days, everybody is adding some type of AI to their product right now. And I would say a lot of those things are not very useful. Right. They just kind of slapped on as an afterthought. And with you guys. Yeah, I mean, you have access to codecs or Claude code or whatever to be able to build more stuff if you want to for the AI. But you don't have a huge team, you don't have, you know, a huge amount of money in the bank that you can just, you know, endlessly build stuff with. So what do you think you did differently to be able to be successful with this? And you've already described a lot of the AI features in the product. So I think people have got a sense of what you're doing. But to back that up, you know, I think you had told me that once you rolled this out, 80% of your customer support tickets were now getting resolved without any human intervention. So it's actually working and the business is growing. But what do you think you did differently to what everybody else is doing.
- [39:46]
B
Well, with some level of simplicity I can say that there are two types of AI applications on the market. One are AI which are trying to add AI because the investors wants it or they believe the customers wants to have it. And for bigger companies this is a big problem. You know, you mentioned Pendo and some others. They are trying to add some stuff but mostly they are trying to acquire some other companies because they are too big to change the direction of the ship. Right. They are scared. They are really scared. They can be gone in two years. I don't think Pando, but you know the address you read about the Monday, you know, recently and stuff. So the world is changing on the other way. If you, if you Read the link LinkedIn Everybody is now bragging about oh, I have a lead gen AI engine. It will book 20, 20 meetings in one second and it doesn't work at all. I had a hobby that in each of this conversation I ask can you hook me up these three customers which are happy using your product at least 6 months crickets never ever. But you know the people like that, you know, always, you know, it's like comment lead and you will get the playbook. So it's very nice. So we want to be some somewhere in the middle. We want to have AI which is delivering. We are Czech people, you know, very. I would say almost pessimistic or realistic. And so we always took care about some meaning in AI. Why to have AI? What problems are we solving? Never have just AI for AI. Just dream about what's possible. Why to put together churn survey manual when you can ask and prompt it. Why to have zillions of people which are doing some discovery calls when you can have it in the application. Of course you can have some AI application which does discovery calls but we have it in our application like part of the application and the rest of the application is reacting on these discovery calls. Like today I was working with product Fruits and I asked Alvin about about or I called him something about integration and after a few seconds he pop up and said, you mentioned integration. Are you still interested? They are there and we have new integration here and here. But unfortunately you would need to to use different tariffs. So you need to. So it was about upsell. So it was amazing how these things are starting together. So I don't think there was something like special or magical. But we always wanted to build AI which we are able to sell because if you are not selling it today, you are disappointing the customers. And I Feel very sorry about their application. I don't know. Lovable address. They are still growing and it's great app, but the churn is massive because a lot of people are disappointed and we can't afford it. We can't have disappointed customers. We can't.
- [43:03]
A
So it sounds like you were a lot more thoughtful about what type of AI features to build and it was less about what can we do with AI and it was more about what do we wish we could have done with this product all those years that we want to do now. What was one feature that you decided not to build with AI that would have been one of those kind of, I don't know, an AI slop feature that somebody might have added on. Because I'm sure when you probably started, you had some of the kind of the obvious things, let's do this, let's do that. Right before you really. You figured out what was worthwhile building.
- [43:47]
B
The initial idea was. One of the initial ideas was to have some AI churn prevention, Churn analytics. Try to. Because everybody's excited, you know, and it didn't work. So it's one of the things he didn't say. Okay, so let's go back to the basic. Let's do the stuff which we, which we have experience with, you know, and our experience is in communication with the user. Communication, you know, everybody is now. And I think it's right. They are betting on this. Agentic systems, independent agenting systems, but they are not really independent. You want to have interaction with them. Right. You want to check what they are doing and stuff, how to communicating with them. This is what we are owning, what we want to own. The communication with application. With AI application. I expect there will be some fluid dashboard. Fluid application which will change based on what you are doing. But still you will need, you know, some visualization, like the narrative that there will be no dashboards. Nothing. It's just wrong. It's just wrong. You have eyes and eyes are the, you know, the contact you want to have. The vision is contact you want to have. If you go to the restaurant, you want to see the pictures of. Of meal, if it's possible, you want to call them and just talk about the food. So the visualization is one of the very important facts even in the times of AI.
- [45:25]
A
Yeah, I think it's super interesting what's happening right now. And I think it sounds like you guys have done a really good job making that shift to using AI in a meaningful way with your product. And in many ways the other advantage that you have is that you're using your product to improve your product, which is a great place to be.
- [45:52]
B
Yes, we are kind of lucky everybody is trying to have AI. I saw toothbrush with AI and real stuff. But we are lucky that we are operating in the area where AI makes a lot of sense. And we are not afraid anymore about this new application which are trying to do the AI onboarding and adoption because they don't have the experience we have. We built over the years talking to thousands of customers. So you need to have some real fundament and build on this fundament, maybe this AI, but you need to have this fundamental.
- [46:35]
A
I have to complain about the toothbrush example you just gave because maybe I'm just a really unlucky guy, but I buy these Sonicare toothbrushes and they always seem to break after one or two years and I have to buy a replacement and a replacement and a replacement. And so when I see them talking about smart features or anything like that, it really makes my blood boil because it's like, I just want you to build a toothbrush that doesn't break.
- [47:14]
B
Right.
- [47:14]
A
That's the big problem. Right. I don't need it to be a chatgpt inside my toothbrush, looking at my teeth and telling me all kinds of shit.
- [47:22]
B
Exactly. I hear you.
- [47:24]
A
If you're building an AI agent, a SaaS product, or stuck trying to scale, check out Gearhart. They can act as your fractional CTO and technical team, bringing AI expertise from projects for Meta and Google plus, strong Silicon Valley connections with founders and VCs. And since they're a Ukrainian born company, you get senior engineers with an offshore pricing model with offices in San Francisco and London and a distributed team of 40 experts. They've helped build over 70 successful products. Right now they're offering our listeners the first 20 hours of development for free. Just book a call@GearheartIO. That's GearheartIO if you want your SaaS product to show up in ChatGPT or Google AI overviews. When people ask for recommendations, Respawna can help. They get your brand cited on high authority sites that AI tools actually pull from completely done for you. And Pay per placement. SaaS podcast listeners get $1,000 in credit on a subscribe and save plan. Just mention my name after you sign up. Visit respona.com that's R E S P-O-N A dot com. Right now, there's a number hiding in your SaaS metrics. The exact point where your growth will stall. You can change it but first you have to see it. It's simple math. When your churn catches up to your new revenue, you stop growing. You're running just to stand still. I've built a free calculator that shows you exactly where your ceiling is and the fastest way to break through it. Find your ceiling at SasClub IO calculator. That's SasClub IO calculator. All right, let's wrap up. Let's get onto the lightning round. I've got seven quick fire questions for you. So ready to go?
- [49:09]
B
Yeah.
- [49:10]
A
What's one piece of startup advice commonly given that you think is wrong?
- [49:16]
B
Talk to customers.
- [49:18]
A
Why?
- [49:19]
B
Because I don't want to use the metaphor with horses and cars, but to large extent it's true. You know, to some extent you also actually want to avoid talking to customers in certain areas when you are in certain flows and you want to just dream about the future. What will work look like? Because this is your domain. Our domain is adoption and onboarding. When I will talk to the people from universities or car building companies, they are not expert. They don't know what they can expect. So working with, like working too much with customers will give you just some small improvements, but not necessary. You will see the longer vision.
- [50:10]
A
What's a book you read recently that stuck with you?
- [50:14]
B
It wasn't recently, but I remember the book from the guy from HubSpot Sales acceleration formula. I think it was from CRO of HubSpot. It's still great book. How to build a sales department. I remember I was, we were struggling with. I wanted to see more yearly deals, right? Not just monthly deals. And our sales guy told me it's not possible. We can't sell yearly deals. And then I read the book and Mark said you can really steer the company just by bonus structure in sales team. So we decided, okay, so from now you have bonus only from yearly deals. And now we have a lot of yearly deals. I don't know how it's happening. It's a mystery.
- [51:07]
A
It's great. What's something you're good at now that you were terrible at in the first year?
- [51:14]
B
Confidence. I didn't have the confidence. I know because I am very, I wouldn't say pessimistic, but you know, I know that we are just like people, just swarm of confused ants hurtling through the cosmic darkness on a chunk of rock that is actually burning inside. So it's, it's crazy. You know what we are trying to do and we are trying to entertain ourselves and, and I wasn't sure that we can really succeed because statistically we were doomed to fail. And I knew it because I'm very realistic guy. I knew it. I was scared. And you don't see the results. It's hard to tell people you have to do this because I know, I do know now when we have something, what we proven, it's easier to convince somebody to do something that it used to be.
- [52:22]
A
What's something you do that most founders might think is a waste of time?
- [52:26]
B
I don't know. The best ideas are coming to me when I am like offline and I really don't think about the work at all, which is uneasy for me. It's really uneasy because I'm obsessed with the ideas all the time. But if I am able not to think about the about the work for two days after these two days, the best ideas are coming to my mind. But probably that's not that unique.
- [52:58]
A
You'd be surprised if you had the time. What's a business that you would start?
- [53:04]
B
I am 54. I don't think I have that many things in front of, you know, see pessimistic guy, swarm of confused ants on piece of rock. But I have an idea which I am playing this because I am also trying to be painter, like oil oil painting and stuff. And I had idea it's not business idea, but when you see the, the paintings of old masters, 2, 200 years old and stuff, you know, in the museums, they look very different from how they looked when the painting was done. Right? It's very different. And I would like to create some twins of these old masters paintings in the way how they looked when these paintings were made. Because with some time and research, I know which kind of colors age and how much and how the colors transition from green to blue and stuff. So that's what I would like to do.
- [54:13]
A
Cool idea. What's something about you that most people
- [54:15]
B
don't know that most of my life I spend making games for women.
- [54:21]
A
And finally, what's something that you're obsessed with that has nothing to do with startups?
- [54:27]
B
Oil painting. That's what I like. I learned I tried to land oil painting when I was like 45. Everybody told me it's not possible. And I think this is what I would like to do after Product Fruits.
- [54:43]
A
Love it. Well, Karen, thank you so much for joining me. It's been an absolute pleasure. If people want to check out Product Fruits, they can go to productfruits.com and if they want to get in touch with you, what's the best way for them to do that.
- [54:57]
B
Email, probably. I'm still email guy. I have a LinkedIn, of course, so they can find me on the LinkedIn and I'm trying not to post post there because whenever I post something, investor are calling me. You can't say this, you can't say this. Delete it. So link it in.
- [55:11]
A
All right, we'll include a link in the show notes. Thank you so much. It's been a pleasure and I wish you and the team the best of success.
- [55:18]
B
Thank you. Bye, bye, bye.
- [55:20]
A
Cheers.