Young scientists and students of the University of Tyumen have created a suite of AI services for promptly informing residents during the flood-risk period. This was reported by the TyumGU press service on March 20. The development includes three products: a Telegram bot with a multi-agent RAG system, a web chatbot, and the "Pavodok72" mobile application for Android. All services use official sources — the region's geoportal, the "Pavodki72.rf" portal, the EMERCOM website, and municipal data. According to the developers, this eliminates neural network "hallucinations".
What problems the service solves
Information about the flood situation had previously been scattered across different sources, while hotlines were overwhelmed with calls. Yevgeny Grinkov, head of the population protection department of the Civil Protection Department of the Tyumen Region, confirmed that the implementation of such services would significantly reduce the burden on call centers and improve the speed of informing citizens.
The Telegram bot answers questions 24/7 about water levels, evacuation points, payments, and rules of conduct. Using the user's geolocation, it builds a route to the nearest temporary accommodation center. The web chatbot is integrated with the Tyumen Region geoportal — it shows up-to-date data from hydrological posts, closed roads, and personalized recommendations. The "Pavodok72" mobile application works offline with a voice interface and an interactive map of safe evacuation routes.
Why the development appeared right now
Roshydromet forecasts water levels rising above normal on the Irtysh and Ishim rivers.
The developers plan to scale the service to other regions of Russia and adapt it for other emergency situations.
Read more materials on the topic:
- Drones of NPTs \"Protos\" delivered medicines to villages near Ryazan that may be cut off by flooding
- An aviation expert explained Russia's need for \"Baikal\": new infrastructure will not be required
- Yamal rescuers deployed \"Orlan-10\" unmanned aerial vehicles to monitor floods