Desti: AI Travel App That Users Loved But Never Booked Through
TL;DR: Desti, a spin-out from SRI International founded in 2011, built an AI-powered iPad app that helped users plan trips by finding destinations based on their preferences. The semantic search technology would analyze reviews and data online to recommend hotels, restaurants, and attractions. The product worked well technically - users appreciated the location suggestions and found the app useful. However, a fatal gap emerged: when it came time to actually book, users went to trusted platforms like TripAdvisor and Lonely Planet. Desti had solved the discovery problem but couldn't capture the transaction. This discovery-to-transaction gap meant Desti generated no revenue despite user engagement. When Series A fundraising came around, investors couldn't justify funding a product with no revenue model. The startup raised $2M across two rounds but couldn't progress further. The story has a partial silver lining: Nokia acquired Desti in 2014 specifically for its semantic search AI technology to enhance Nokia's maps platform. The team members joined Nokia, preserving their employment and the technology investment. However, the Desti product itself was shut down shortly after acquisition. This case illustrates both the danger of solving discovery without owning transaction, and how valuable technology can survive even when the business fails.
Key Insights
- Users loving your product doesn't mean they'll pay through your product - Desti users appreciated recommendations but booked elsewhere
- Entering a market with established giants like TripAdvisor requires users to abandon trusted brands, which rarely happens
- Revenue model must be validated early - user engagement without monetization path is not fundable
- Valuable underlying technology can be acquired even when the product fails (acqui-hire)
- Being a discovery layer without owning the transaction is a losing position in travel
Actionable Takeaways
- Validate the full user journey from discovery to transaction, not just the discovery experience
- Test whether users will complete the revenue-generating action through your product, not just whether they find it useful
- In markets with established giants, find a niche they don't serve rather than competing head-on
- Build revenue generation into early product versions to prove the business model before Series A
- If product fails but technology is valuable, consider acqui-hire as an exit path that preserves team employment