Speaking at App Promotion Summit London 2025, Anton Volovyk reveals how his company went from viral success to spectacular failure — and back again.
The golden age of easy app growth is over, and artificial intelligence is both the problem and the solution. That’s the stark message from Anton Volovyk, co-CEO of Reface, the AI face-swap app that once dominated download charts with 250 million installs before having to grapple with the consequences of a major product failure that brought its momentum to a halt.
Speaking at App Promotion Summit London 2025, Volovyk laid out a brutally honest timeline of his company’s evolution — from accidental viral sensation to humbling market lesson to strategic AI app portfolio — that serves as a masterclass in surviving the new reality of consumer app growth.
The five-second rule that changed everything
Reface’s original success came down to one critical insight: users will tolerate exactly five seconds of wait time for AI-generated content, but not forty. When the company launched its follow-up app Restyle in 2022 — a video-to-video filter technology that Volovyk describes as “very innovative” for its time — that seemingly minor difference killed the product.
“We thought, okay, maybe it’s not a big problem, but we are in the market of fast AI entertainment,” Volovyk explained. “It’s like Haribo bears at the supermarket counter — an instant purchase. If a person needs to wait 40-60 seconds, it’s so much worse than five seconds, it’s crazy.”
The failure was particularly stinging because Restyle had all the hallmarks of success: impressive click-through rates on ads, celebrity partnerships (including Jared Leto), and cutting-edge technology that made Reface only the second company after Runway ML to offer consumer-focused video-to-video functionality. But users didn’t stick around.
“If we like something, it doesn’t mean the market likes it,” Volovyk admitted, describing how Reface “bet the entire company” on Restyle, pulling resources from their successful flagship app for what turned out to be a costly lesson in product-market fit.
Growing and scaling AI apps
Source: App Promotion Summit
The new playbook for AI app success
That failure forced Reface to develop what Volovyk calls a more “normalized” approach to building AI consumer apps, one that treats artificial intelligence as a commodity rather than a competitive moat. The company now operates as an AI app studio with over ten products including a gamified food tracker (BitePal) and a tattoo design generator that hit #1 in the U.S. App Store’s Graphics & Design category.
The secret? Three rules that flip conventional wisdom about app development:
Build for the TikTok attention span
Apps must capture users within 1-3 seconds or lose them entirely. This means creating memorable mascots, distinctive visual elements, and instantly recognizable branding that can work across dozens of ad variations. BitePal’s raccoon character, for instance, is great for building brand recognition and can be used to easily generate multiple creative assets.
Build a creative testing lab
The most successful apps aren’t guessing what users want, they’re systematically discovering it through rapid experimentation. Reface treats each ad creative as a data point, using performance metrics to uncover unexpected user segments and pain points. Their approach: create vastly more content than competitors, fail faster, and let the data reveal which audiences actually convert.
Freemium still works, but differently
“We force ourselves with freemium to make actually good products so people use them knowing what they’re buying,” Volovyk explained. Rather than hard paywalls, successful apps let users experience core value before converting. This approach forces teams to build genuinely useful products while creating higher-quality subscribers who understand what they’re purchasing. The results are better retention and fewer refund requests.
The end of the AI monopoly
The broader challenge facing AI app developers is that their technology advantages are evaporating. What once required specialized teams and proprietary models can now be built using open-source tools.
“Two or three years ago, we were one of the few teams in the world that could create these kinds of models. Now it’s all open source,” Volovyk noted. “The AI revolution happened, and for us it was a big negative.”
This shift is forcing companies like Reface to compete on execution rather than innovation, treating AI as infrastructure rather than product differentiator. The winners will be those who can combine AI capabilities with deep understanding of user behavior, sophisticated marketing operations, and sustainable business models.
In the age of AI abundance, the apps that win will be the ones that master the decidedly un-sexy fundamentals of user acquisition and retention.
As Volovyk put it: “Some fundamentals are the same.” The difference now is that everyone has access to the same AI tools, so execution is everything.
