Building ZAN: A Decision Engine for Modern City Life
Every product starts with a frustration. For ZAN, it was the ten minutes we spent every single day arguing about where to eat lunch โ cycling through the same apps, seeing the same listings, eventually giving up and defaulting to the nearest familiar spot.
The insight wasn't that we needed more options. We needed a faster path to a good-enough answer. The kind of answer a local friend gives you when you ask: 'What's good around here right now?' They don't hand you a spreadsheet. They give you one or two picks with a short reason why.
The first version of ZAN was barely an algorithm. It was a ranked list updated manually every morning by one person who knew the area well. Users loved it precisely because of the constraint โ not 200 options, just five, curated for today.
From there, the challenge became: how do you automate the intuition of someone who knows the city? That question drove our engineering roadmap. We integrated traffic APIs, weather data, local social signals, and order velocity patterns into a scoring model that updates continuously.
The name ZAN comes from the Vietnamese 'ฤn' โ to eat. But the vision was always bigger than food. Eat, Go, Move: a decision layer that starts with the most basic daily question and expands into how people navigate and experience their city. Food was the right entry point because it's universal, high-frequency, and time-sensitive. If we can solve it well here, the same decision engine applies to dozens of adjacent problems.