AI is democratizing many processes, from access to knowledge to software development. At Vestbee, we're increasingly seeing solo founders with no technical background pitch us working MVPs they built using AI. Prototypes, landing pages, market research, first-draft go-to-market strategies — all of it can now be reduced to prompts.
When execution becomes cheaper and more accessible, it stops being a meaningful differentiator. At the pre-seed and seed stages, investors increasingly rely on something harder to quantify: the founder's ability to learn quickly, adapt, make decisions under uncertainty, and recover when things don't go according to plan.
That shift comes at a time when the European fundraising environment is becoming harder to navigate and more concentrated: as Vestbee reports, in H1 2026 growth-stage companies accounted for 72% of funding, while early-stage startups, despite representing more than 90% of all deals, saw their share of total capital fall to just 28%. Capital is also moving between sectors — in 2020, marketplaces and consumer startups looked like safe bets, but nowadays investor appetite is increasingly moving towards hardware. Even strong execution cannot compensate for changing market dynamics.
Even AI is no longer a safe haven of infinite growth. It continues to attract talent, customers and capital at an extraordinary pace, yet we're already seeing the first wave of AI startups shutting down.
You wouldn't know that from reading most startup headlines or scrolling through LinkedIn, where success stories dominate the conversation. They are polished, compressed, and often told as if the path from idea to product was almost inevitable. The years of false starts, abandoned ideas and difficult strategic decisions rarely make it into the final narrative.
The goal of this article is to look behind those polished stories and examine the biggest lessons learnt along the way: companies that changed direction, sometimes more than once, before finding product-market fit, as well as startups whose pivots ultimately weren't enough to keep them alive.
One thing is worth stating at the outset — pivot stories are almost always simplified and cleaned-up narratives told in retrospect. Reality is much messier. Even so, these stories are valuable because they provide a more realistic benchmark for founders — successful companies are often built through multiple attempts, and finding the right direction can take much longer than many of them would initially imagine. Here are four European startups, at different stages of their journeys — from early-stage lessons, to international successes, eventually to failure- worth looking into.
LAM’ON: everything takes much longer than initially expected
Angela Ivanova and Gergana Stancheva discovered a major sustainability problem while working in Bulgaria's printing industry: plastic laminates made paper unrecyclable. Unable to find a biodegradable alternative, they decided to develop one themselves. LAM’ON is still an early-stage company finding its footing in a much larger market, but its journey already offers a clear example of how a startup can evolve. CEO Angela Ivanova shared the lessons behind that evolution with Vestbee.
"At the beginning, we thought we were developing a product," she told Vestbee. "In reality, we were building a materials company." That realization led to the company's first commercial pivot.
What began as biodegradable laminating film for the printing industry expanded into flexible packaging, opening a much larger market. Today LAM'ON develops bio-based films for e-commerce and industrial packaging, shrink film, and increasingly food packaging.
For Ivanova, however, the commercial pivot was the easier one. "The harder pivot was in our own mindset. Inventing a material is one challenge. We had to become manufacturers, build supply chains, understand machinery, certification, quality systems, regulation and industrial sales."
The founders knew the material they wanted to create. What they underestimated was everything required to bring that material into an industry built around decades-old manufacturing processes.
"You are not simply asking a customer to buy from a new supplier. You are asking an industry to trust something it hasn't used before. That means testing. Then another test. Then a different machine, another formulation, another certification, another procurement department."
That experience fundamentally changed how Ivanova thinks about innovation. "We learned that being technically right isn't enough. The material has to work within the customer's economics, machinery, supply chain and regulatory reality."
LAM'ON was also forced to adjust to a changing funding environment. The company was founded when Europe's green transition was becoming a major investment priority, but the geopolitical landscape has since shifted. "Those priorities haven't disappeared, but the world around them has changed. Wars, geopolitical instability, energy security and the need to strengthen Europe's defence capabilities have inevitably shifted where governments and investors are focusing capital. Today, climate and bio-based technologies are competing for attention in a much more complex strategic environment."
Rather than repositioning around whichever theme was attracting capital, the company changed its expectations."We cannot build a company on the assumption that capital will be available simply because the technology contributes to Europe's sustainability goals. We have to demonstrate that the business works: customers, margins, production efficiency, scalability and a credible path to profitability."
That shift also changed Ivanova's own definition of success.
"I am much less impressed by announcements than I used to be. Raising a round is great, winning a programme is great, getting recognition is great—but none of those things mean that you've built a sustainable company."
Asked whether AI has changed the business, her answer is immediate. "AI hasn't changed the fundamental problem we are solving. You cannot prompt your way out of manufacturing a physical material."
If anything, she says, AI has sharpened the distinction between digital and industrial businesses.
Today, the challenge is no longer inventing the material. It is building the company capable of producing it at industrial scale."We spent years learning how to make the material. Now we have to build a company capable of producing a lot of it."
Synesthesia: finding the market before chasing the vision
A company’s earliest stages are often the most turbulent ones, and Synesthesia is a great example. The company was founded in 2017 with an ambitious goal: making it possible for anyone to create Hollywood-quality films using AI. Nowadays, it is a reality in the making, but almost 10 years ago no one was ready for such innovation. Instead of waiting for market conditions that would accommodate the vision, Synthesia looked for a problem they could solve with the tools they already had. The company pivoted into AI dubbing, translating videos into multiple languages while synchronizing voices and lip movements for film studios and advertising agencies.
"The technology didn't really work, and everybody thought we were completely crazy," CEO Victor Riparbelli told Sifted.
The business generated revenue, but it exposed another problem. Every customer required custom work, pushing Synthesia toward becoming a visual-effects agency rather than a software company. Its defining pivot came next: from AI dubbing to AI avatars for enterprise communication. Filmmaking was the original use case, but what proved to be scalable was much less glamorous — corporate training, compliance and internal communications, where companies needed thousands of videos localized across multiple languages rather than one perfectly crafted production.
The underlying technology changed far less than the business model. What changed was where the technology created value. By prioritizing practical utility over technical novelty, Synthesia transformed itself from an experimental AI video startup into an enterprise software company valued at $4 billion following its Series E round in 2026.
n8n: jumping on the AI bandwagon early
n8n is another interesting example of early-stage pivots with AI at its core. The company started as a low-code workflow automation platform, and by 2022 it had a strong product, an enterprise customer base and a community of more than 16,000 users, but growth was steady rather than exceptional.
The turning point came in 2022, months before ChatGPT launched, as founder Jan Oberhauser was watching companies such as Pinecone narrow their positioning around AI and began thinking about what the shift could mean for n8n. “Before they were a vector database, which was cool. Then they became an AI database, which was very cool,” Oberhouser said to Matt Quinn.
AI wasn't even on the company's roadmap, but he soon realized it could threaten some of n8n's existing use cases by automating the simpler workflows users relied on. So he changed course, and n8n became a company we know it to be today: a platform for building AI applications.
But the initial launch fell flat, as users feared that n8n was replacing the product they knew rather than extending it. The transition made sense business-wise, but the customer base proved to be resistant. For nearly a year, the company worked to bring its community along, explaining the new direction, improving the product, and showing users how AI could extend the workflows they were already building. By the end of 2023, Oberhauser effectively had to launch the product for a second time.
The results changed the scale of the business. By 2025, n8n's user base had grown sixfold and revenue tenfold, with more than 80% of workflows involving AI agents. Oberhouser concluded that AI could have eaten into n8n's business within a few years if the company had waited for the market to mature. By May 2026, n8n was valued at $5.2 billion following a strategic investment from SAP.
Quesma: when everything looks right, but the startup still fails
Researching this piece was not easy, as we wanted to include early-stage lessons, successful pivots and market adaptations, but also companies that ultimately didn't make it. Those stories are far less common and, when they are told, usually appear as a footnote or a brief post-mortem. Quesma CEO Jacek Migdal offered something different: a candid account of his company's closure, including the decisions that led there and what he learned from them.
Migdal identified a problem: database migrations were difficult, and a proxy layer could make them easier, and raised $2.5 million to build a database gateway. Potential customers were interested, praised the technology, but too few actually ran the product or committed to pilots.
That gap eventually became impossible to ignore. The MVP took nine months instead of the planned five, and by the time it was ready, the market was narrower than the team had assumed, as the potential customer base started to look like large, established organisations that moved slowly.
Quesma found itself in what Migdal calls "pilot purgatory": plenty of interest, very little urgency. After analysing lost deals, the team realised that Quesma was often around the twelfth priority for potential customers. In companies that completed only two or three major initiatives each quarter, being interesting was not enough. The company also discovered that customers often saw Quesma as a feature of a broader data solution rather than a standalone product. Meanwhile, AI coding assistants were making database migrations faster and cheaper, weakening one of the assumptions behind the original business.
Rather than continue chasing pilots in a shrinking market, the company accepted that its technology had more value as part of someone else's platform than as a standalone business. Selling its IP wasn’t the original plan, but Quesma eventually sold it to Hydrolix, its largest customer, giving the team enough runway to pursue its next idea. What Migdal showed was that startups are often redirected by the accumulation of small signals: slower-than-expected product development, changing markets, weak customer urgency, evolving technology, and difficult founder decisions.
Taken together, these stories show that pivots are rarely one huge and often dramatic decision. More often, they are responses to a set of patterns: customers behave differently than expected, markets shift, technology changes, or a business model proves impossible to scale. The real test is whether the founders can recognize those signals early enough and decide what to do with them.
