Show notes
You can build faster than ever with AI.
That does not mean users will adopt what you build.
In this episode of Marketing Spark, I talk with Karel Papik, co-founder and CEO of ProductFruits, about one of the biggest challenges facing SaaS companies today: the growing disconnect between product velocity and user adoption.
We explore why companies are shipping more features while users feel increasingly overwhelmed, why many AI tools struggle to deliver lasting value, and how onboarding has evolved from static product tours into personalized AI-driven experiences.
Karel also shares:
- Why AI products often suffer from weak retention
- The hidden friction slowing SaaS adoption
- Why companies hesitate to switch software despite dissatisfaction
- How ProductFruits competes against larger players like Pendo and WalkMe
- Why smaller SaaS companies can still win in crowded markets
- His perspective on the future of AI, product experience, and human work
This conversation is especially relevant for SaaS CEOs, founders, marketers, and product leaders trying to understand why growth becomes harder even when teams are building and shipping faster than ever.
From the conversation
Why AI products struggle with retention
Karel Papik, co-founder and CEO of Product Fruits, does not think the main issue is how fast companies ship. He sees a quality problem. He compares it to a town full of good restaurants where some begin serving poor food, until people stop eating out at any of them, including the older ones.
Karel points to retention numbers as evidence. He believes median annual retention for AI tools is around 40%, against roughly 90% for traditional B2B SaaS. He says those figures look more like B2C SaaS. He also notes that Lovable reportedly has 85% retention at day three, but he would like to see the nine-month figure.
"I think the problem is not in the pace, but in the quality of the new features and new platforms and new applications are being shipped."
AI-personalized outreach that adds no value
Karel offers cold email as an example of AI that has not delivered. Two years ago he received two or three cold emails a day. Now he gets about seven, even with a spam filter running.
The messages are personalized by AI, but Karel finds the personalization obviously fake and annoying, and he does not respond. In his view, AI has made this kind of outreach cheaper to send without making it more valuable.
"But the question is if the value is really there, the stickiness and the retention of AI tools and platforms is actually significantly lower than traditional B2B SaaS."
Why companies hesitate to switch software
Mark raised the idea that buyers are stalling because there are so many options, and doing nothing feels safest. Karel agrees. He says many companies sense a large change coming and worry about whether their own product, company and jobs will stay relevant. That makes them hesitate over any new tool.
Product Fruits responds by keeping its classic tools, such as step-by-step tours, hints and video feedback, alongside its newer AI features. Karel says the company does not force anyone to adopt the AI. He believes customers buy licenses because they see Product Fruits as a path toward the future.
Barriers to adopting AI tools: uncertainty, pricing and the black box
Karel names uncertainty as the biggest roadblock. A calculator always gives the same answer, while AI output is fuzzy, so users have to accept iteration. Large companies are wary because they cannot be fully sure what the AI will produce.
Pricing is another barrier. Product Fruits tried charging per resolution about six to eight months ago, which seemed fair because customers paid only when the product helped. Buyers still wanted a fixed number to take to procurement. The company now offers yearly and monthly tariffs and pays its AI providers per resolution, so the risk sits with Product Fruits.
Karel also says companies do not want a black box. Product Fruits spends considerable resources on governance policies to keep its AI on a short leash.
From product tours to personalized onboarding
Product Fruits began as a digital adoption platform built around step-by-step product tours. Karel now calls it a product experience platform. He says AI finally made possible things the team had long wanted to do.
In the old model, the application owned the onboarding and told users what to click. Karel says people do not want that and prefer to decide for themselves with some help. Product Fruits now uses a question agent that holds a short discovery conversation about the user's use case, experience with similar platforms and department. An onboarding agent then tailors the experience based on those answers.
The same data produces outcomes for admins. They can ask where users struggle most and what they dislike, and use the answers to improve the product.
"Because when AI came, we had finally the option to do the stuff which we always dreamt about."
How a 26-person company competes and sells
Product Fruits has 26 people and sells worldwide, with the US as its biggest market. Karel names AI as the edge over larger players. He says competitors are trying to change course but moving slowly, while Product Fruits can be more agile. Marketing is a standard mix of content and PPC, run by Dan and Lucas, with no events. Karel says there is no silver bullet in lead generation, so the company focuses on the product.
The company has about 300 companies in free trial at any time. It once converted 22 to 23% of them, and now converts below 15%, but the accounts are much bigger. Karel says Product Fruits rebuilt its sales team three times, and the third attempt is working. All leads are inbound. Salespeople have to broaden what buyers ask for, and the company is upselling AI to its 1,300 customers.
Karel adds that the company is profitable, has raised $4.5 million in seed funding and is not pressed for runway.
Karel's view on AI, jobs and the future of product experience
Karel admits the long-term outlook for AI and human work is grim, but expects the next twenty years to bring benefits. In the nearer term, he doubts fully headless applications are the answer, because people need control, governance and iteration, and that requires an interface.
He also does not see AI replacing people who enjoy their jobs. He uses three AIs at once to work faster and better, and sells Product Fruits as cooperation with users, not replacement.
His vision is an invisible companion that sits beside the user, helps inside the application and holds knowledge of the user's field.
"I don't believe much in these headless application where the agents will just roam around and do everything on their own."