Yadro has invested about two hundred million rubles in the development of the MDT Vision system - a neural network that analyzes the quality of mobile communications based on anonymized data from users' smartphones. In essence, it replaces classic "field tests" when operators manually check the network.
AI sees what used to be a "blind spot" - for example, the quality of communication inside buildings or in difficult urban conditions. The system collects data on coverage, load and speed, and then completes the missing picture and helps operators to more accurately configure the network.
According to the test results, the effect is already noticeable: the network capacity has increased up to 30% without installing new equipment, the speed for users has increased by 20-40% percent, and the number of on-site inspections has been reduced by 70-90%. In some regions, this even reduced subscriber churn by 15-20%.
The idea itself is not new - operators have been collecting such data for a long time. But previously, the accuracy left much to be desired: classic methods gave an error of up to 200-300 meters. With the use of AI, it was possible to reduce it to about 50-100 meters.
This opens up new opportunities: operators can more accurately configure antennas, redistribute load and improve coverage without large-scale investments in infrastructure.
At the same time, the market remains cautious. Telecom companies are in no hurry to implement new solutions without proven effectiveness in practice. Even a strong product can go through a long pilot phase before a full launch.
But the trend is obvious: communication is beginning to improve not only with hardware, but also with algorithms. And in some cases, this is already enough to significantly speed up the Internet without installing new base stations.
