AI Agent in Sirius Can Answer Bus Schedule Questions and Recommend Things to Do
Two Russian digital technologies for the urban environment were unveiled in the federal territory of Sirius. Videofor-Yug demonstrated a situational AI agent, while Eksta presented Neurodetector, a traffic analysis system. The technologies were showcased at a meeting of the Innovation Science and Technology Center community.
Both projects apply AI to real-world urban challenges. Computer vision will collect traffic data, while conversational AI will answer questions from residents and tourists.
At the Bus Stop and the Intersection
A conventional chatbot follows a simple question-and-answer model. It does not know where the user is, what time of day it is or what is happening nearby. A situational agent works differently. It is tied to a specific location, the Sochi Park bus stop, and has access to information about nearby venues and events, as well as the transit routes serving the area. It can readily suggest, for example, where to go with a child or when the next bus will arrive.
The agent offers several modes: navigation, transit information, an events calendar, safety information and recommendations for activities with children. This AI assistant is not just knowledgeable; it can also take the initiative. Unlike a conventional chatbot, it can start a conversation by greeting users and offering to tell them about the day's events. The pilot is currently being tested at a single bus stop, but the developers plan to build a network of such locations.
Another project, Neurodetector, uses a smart camera to turn video footage into structured numerical data about traffic. The system counts passing vehicles, measures their speeds and assesses traffic volume. It can analyze traffic in different weather and lighting conditions. Experts can use the resulting data to identify trouble spots and plan traffic-management measures. Developers have selected the intersection of Mezhdunarodnaya and Chempionov streets as the test site for the video analytics system, a challenging location with multiple approaches, turns and intersecting traffic flows.
I Asked … Ann and Alexandra
The closest counterpart to Sirius's situational agent is Alexandra, Moscow's public transit digital assistant, which operates across apps, websites and social media. In five years, Alexandra has answered more than 17 million questions and now handles over 24,000 requests per working day. By 2030, that figure is expected to reach 35,000 to 40,000.
In 2025, two large language models were integrated into Alexandra. She can now better understand what users mean, remember the context of a conversation and sustain a dialogue. Interactive kiosks called Live Communication, featuring an Alexandra avatar, are installed at Maryina Roshcha metro station and the Nizhegorodskaya urban rail terminal. The avatar provides directions and information. The approach is part of an international trend. In the US, Ann Arbor has its own conversational AI assistant: AskAnn, which helps residents find information about municipal services in 71 languages.
Meanwhile, Voronezh has integrated VisionLabs' LUNA CARS video analytics system into its intelligent transportation system. It processes data from more than 50 traffic controllers and 100 cameras. With accuracy exceeding 95%, it identifies license plates, vehicle makes, models, colors and types. It also counts vehicles to detect rush-hour congestion and recognizes traffic violations, including crossing a solid line, making an illegal U-turn and parking where it is prohibited. The system can also detect people in restricted areas and animals on roads. In Bern, Switzerland, officials have also begun deploying YutrafficawareAI, a system that uses video analytics to manage traffic signals and recognize pedestrians and cyclists.
Green Light for Technology
By the end of 2025, 32 Russian urban agglomerations had reported reaching at least Level 1 of maturity in their intelligent transportation systems. The federal budget allocated 2.7 billion rubles (approximately $34 million) to their development that year. The infrastructure needed to support solutions of this kind is already taking shape.
Sirius is serving as a test bed for Russian digital technologies designed to make urban information easier to access and improve traffic-flow analysis. The projects' future will depend on pilot results, their potential to scale and demand from city operators. Once their accuracy, reliability, user-friendliness and compatibility with urban infrastructure have been confirmed, other regions will be able to adopt them.
