Milena Nikolić has a specific complaint about artificial intelligence, one that's been gnawing at her since her days running engineering at Trainline. It's not about hallucinations or bias or any of the usual concerns. It's simpler, almost mundane: AI makes people read too much.
"People want structured, visual, and interactive information," she told TechFundingNews back in February. "Walls of text degrade the experience."
This might sound like a minor gripe—the kind of thing a designer obsesses over while the rest of us just scroll past another ChatGPT essay. But Nikolić isn't just complaining. On February 12, her London-based startup Heywa Labs announced it had raised $5 million to prove her point, backed by Cherry Ventures and a cluster of early-stage investors betting that the future of AI search looks less like a conversation and more like flipping through Instagram Stories.
The premise: replace ChatGPT's paragraph problem with something the company calls "Generative UX"—interactive cards that rearrange themselves based on what you're actually trying to do. Ask about weekend recipes, get tappable cards. Plan a trip, navigate a timeline. Look for workout routines, see a comparison table materialize.
It's an ambitious play in a space where Google, Microsoft, and a parade of well-funded startups are all racing to crack the same puzzle. The question isn't whether AI interfaces need fixing—most people seem to agree they do. The question is whether a seed-stage team can actually build something better.
Google pedigree, Pinterest vibes
Nikolić didn't leave Google alone. Her co-founder, Martín Sturla, came from the same hallways—she'd directed engineering for Google Play while he worked on AI Search and Play Integrity. Both had watched the company grapple with how to present AI-generated answers, and both walked away convinced the current approaches were missing something fundamental about how people browse on their phones.
Cherry Ventures partner Christian Meermann summed up the pitch in a LinkedIn post as "ChatGPT meets Pinterest," which is either reductive or brilliantly concise depending on your tolerance for venture capital shorthand. The round also pulled in Plug & Play, OpenseedVC, Pareto Holdings, and Ventures Together—a typical constellation for a European seed deal in 2026.
The app itself has been live since December 10, quietly available on iOS in the US, UK, and Canada. But the funding announcement represents Heywa's first real attempt to step out of stealth mode and make a case to the broader market.
The manifesto problem
Tech startups love manifestos. Usually they're marketing fluff dressed up as philosophy. Heywa published one at generativeux.com, and it's worth taking seriously—if only because it reveals how the team is thinking about constraints.
The core argument: instead of trying to generate infinite variations of interface elements (buttons, menus, layouts), build a fixed set of components that can reconfigure instantly. Cards, timelines, comparison tables. The system decides which to use based on your query, then streams them in as fast as possible.
Speed matters here more than you might expect. According to Cherry Ventures, Heywa reaches "first meaningful interaction" in roughly two seconds. The manifesto takes a swipe at what it calls adaptive UIs that "take ten seconds" to load—a clear shot at some of the more elaborate generative interface experiments happening elsewhere in the industry.
That speed obsession isn't just about user experience. It's a technical bet that lightweight, pre-built components will always outperform systems trying to generate custom interfaces from scratch. Whether that holds up at scale, with more complex queries and deeper personalization, remains an open question.
Vibes as a filter
Here's where things get interesting, or possibly gimmicky—it depends on your perspective. Heywa lets users filter results "by vibe."
Yes, really.
The idea seems to be that when you're searching for inspiration rather than information—weekend recipes, gift ideas, hobby projects—you don't necessarily know exactly what you want. You just know what feels right. The app pulls from sources including Reddit and TikTok, then lets you browse with an emphasis on aesthetics and mood rather than keyword matching.
It's the kind of feature that either clicks immediately or leaves you wondering whether it's solving a real problem. Early reviews are sparse—just five ratings on the App Store, averaging a perfect 5.0—but one user in January specifically praised the "story format" over "long boring ChatGPT paragraphs." Take that data point for what it's worth.
The everyday use case
Heywa isn't chasing enterprise customers or trying to replace Google for complex research queries. The pitch is explicitly consumer-focused: cooking, travel, fitness, DIY projects. The App Store description positions it as a way to "explore like flipping through stories," leaning into mobile-first, visual browsing behavior that Nikolić's team believes better matches how people actually use their phones.
Since launch, the company has pushed updates aggressively—five by the end of January, including iPad support on the 31st. The team runs somewhere between 10 and 12 people (sources varied slightly), with backgrounds spanning Google, Apple, Snap, and Trainline. Small enough to move fast, experienced enough to know the pitfalls.
The app is free, with no ads. Monetization plans? The company hasn't said publicly, which is either strategic patience or a problem they're still figuring out. Probably both.
A crowded field
Heywa isn't alone in trying to reinvent AI interfaces. Perplexity launched "Pages" to turn searches into structured articles. Microsoft's Copilot Search now blends generative text with richer web cards. Arc Search built a mobile browser that condenses entire queries into single-screen summaries.
What distinguishes Heywa—at least in theory—is the emphasis on "generating experiences" rather than just reformatting outputs. The manifesto makes a distinction that might sound like semantics but matters in practice: instead of changing what the interface looks like, change how it behaves as you interact with it.
Whether that philosophical difference translates into something users actually notice and prefer is the $5 million bet. Incumbents have resources, distribution, and years of user behavior data. Heywa has a thesis, a prototype, and a runway to prove it works.
What comes next
The funding buys time, but not certainty. Nikolić's team is treating Heywa as both a consumer product and what investor materials describe as a "laboratory" for testing Generative UX principles. That dual identity—product and platform—suggests the company might be keeping options open about where this ultimately leads.
For now, the app is live, the concept is real, and the market is paying attention. Whether people actually want to search this way, whether the interface holds up under real-world pressure, whether a seed-stage startup can outmaneuver giants with infinitely deeper pockets—those questions don't have answers yet.
But Nikolić is betting that if you can make AI feel less like reading a term paper and more like browsing a magazine, people will notice the difference. Sometimes the most ambitious ideas start with the simplest frustrations.
