Russian Business Stops "Playing with AI": Now Tools Are Selected Based on Strict Metrics

The key barrier was the lack of personnel capable of competently formulating tasks for neural networks

More than half of Russian companies have already implemented artificial intelligence technologies, but the market has entered a new phase. This was stated by Prime Minister Mikhail Mishustin at CIPR-2026, emphasizing that experiments with AI are giving way to projects with strict business indicators and project management.

The numbers confirm the shift: in the telecom industry, 62% of processes are already automated, and in the banking sector, 80% of currency control operations are performed by neural networks without human intervention with an accuracy of about 90%. However, businesses have encountered an unexpected barrier – people simply don't know how to formulate tasks for AI. With the emergence of a perfect executor capable of generating code in minutes, it turned out that employees who are supposed to set tasks for it do not possess such a skill.

There are still areas where AI will never be allowed. Andrey Korolev, Deputy Director for IT at Rosatom, emphasized that in nuclear energy, the cost of error is too high. Neural networks only assist in design but are not allowed to manage technological processes.

The market is divided into two circuits: in one, AI takes on more and more operations, in the other, it remains a tool under strict human control. The main challenge is not access to technology, but the ability to manage it competently.

Read more on the topic: