KhariBulbul
In Baku and surrounding areas (such as Khirdalan, Sumgayit, and surrounding suburban settlements), navigating the "last mile"—reaching the final destination after exiting a metro station or during heavy traffic on route buses—causes significant time loss. Apps like AYNA MAAS or BakuBus only show standard fixed routes, failing to use AI to optimize for weather conditions, peak rush hours, or overall travel costs.
Preliminary score
- Value for the user15.0 / 25
A specific problem and a clear outcome from solving it.
- Prototype and use of AI11.0 / 30
A working core scenario, and what AI actually contributes to the result.
- Quality testing10.0 / 20
Tests, examples of failures, and comparison with the current approach.
- Feasibility6.0 / 15
Data requirements, running costs and a clear next step.
- Originality4.0 / 10
A meaningful difference from common solutions.
What the judges wrote
- GPT-6 AstraPreliminary50/100
- Strengths
- Your focus on Baku’s last-mile commuters identifies a concrete local problem, and the Azadliq Avenue–Icherisheher example clearly presents the time–fare trade-off. Your testing notes identify actionable failures in location handling, localization, small-screen layouts, and identical endpoints rather than relying on generic success claims.
- To improve
- Your prototype evidence should show a complete routing run and explain the runtime AI contribution: listing v0 does not establish AI-based routing, and simulated traffic should be clearly distinguished from live data. Your tests would be stronger with reproducible inputs, expected versus actual outputs, and route-time and fare comparisons against existing apps or real journeys. Your feasibility plan should identify sources for transit schedules, fares, and scooter availability, estimate running costs, and prioritize one achievable integration. Your enterprise-track case also needs a specific business customer, workflow, and measurable adoption outcome.
- Claude Opus 5.5Preliminary42/100
- Strengths
- Your problem framing is grounded in local reality: last-mile trips from metro stations and suburbs like Khirdalan and Sumgayit, plus the gap left by AYNA and BakuBus showing only fixed routes. Your interface compares taxi, metro-plus-walking and scooter-plus-metro options with time and price side by side, which is easy for a commuter to act on. Your testing section is refreshingly honest, listing concrete failures such as incomplete AZ/EN/RU translations, hangs on custom location names, layout overlap on small screens, and identical origin and destination not being rejected.
- To improve
- Your materials list v0 as the only model, which suggests AI helped build the app but does not run inside it. 'AI Combined Routing' and 'Live Traffic Simulation' need to show what model or learned logic produces the recommendation, rather than fixed rules or simulated values. Back the '45+ minutes lost daily' claim with a source, and measure actual time or cost saved on a set of real Baku trips compared with AYNA or Google Maps. Add running costs, a data plan for real bus timetables and traffic, and frame a buyer and adoption metric that fits the Enterprise track, for example a city transport operator or employer commute programme.
The team
İsmayıl ÖmərovLead
product
Niyazi Musayev
developer
Ali Muradov
developer