Search has a bias built into the result.

Search-based discovery has a bias built into the result. The most visible places are the ones that paid for visibility, or the ones that got SEO right — neither of which correlates with what's actually worth your attention nearby. Urbee's bet was simple: most of the time, you're not looking for something specific. You're looking for what's around you, organized clearly enough to choose. And open data, properly aggregated, was rich enough to power that experience without a single sponsored slot.

The product was launched in March 2017 with one pilot city.

Two faces, one map of the city.

Urbee aggregated open municipal data — POIs, events, public services, transport, local information — and re-presented it as a visual, location-aware discovery surface. Tap a category, see what's near. No keywords, no filters to configure first. Beyond the consumer side, the product had a second face for local authorities: a way for municipalities and tourism offices to push contextual information directly to people physically present in their territory. The right notice, to the right person, at the right place.

A product that worked, killed by a business that didn't.

Three months after launch, retention from the early-adopter cohort in our pilot city sat around 30%. Modest in absolute numbers — a few hundred users at peak — but a strong signal that the proposition worked. People came back to discover their city, not just to use the app once.

The problem wasn't the product. It was the customer. Selling to local authorities meant 12-to-18 month sales cycles, payment terms measured in quarters, and a procurement process that didn't match the cash needs of an early-stage startup. We could land municipalities, but we couldn't survive the time it took to get paid. In 2018, we wound the company down.

The instinct held up. The business model didn't.

What Urbee tried to do in 2017 — replace the search bar with contextual discovery — is now a mainstream pattern. LLM-powered apps (Perplexity, Arc Search, the entire AI-search wave) are exploring the same question we asked: what would a discovery experience look like if you didn't have to know what to look for? The product instinct held up. What I'd change is the business model. B2G was the wrong shape of customer for an early-stage startup's cash cycle, however legitimate the buyer.