Trend-Group and Sber presented an AI agent project at SberPro Tech 2026

On October 30, 2026, at the V conference on digital business transformation SberPro Tech 2026, the Krasnodar developer Trend-Group and Sber presented a joint project on creating and managing AI agents. This was reported by the primary source.

An AI agent is a program that receives a task, independently chooses a sequence of actions, and calls the necessary tools: email, a database, internal services. What distinguishes it from a chatbot is that it executes a chain of steps rather than answering with a single reply. Agent management is a separate part of the project: one needs to set rules for them, track results, and restrict access to systems.

How this differs from the previous approach, the source does not specify. Usually, before agents, routine operations are handled by regulations and manual work in several systems, but there is no confirmation of this in the note. There is also no data on which specific tasks were automated at Trend-Group.

The only figure from the primary source is the time spent on routine tasks: three hours per week instead of 14. This is a claimed saving, not a measured methodology. How the hours were counted, over what period, and on which group of employees—there is no data. From these numbers one cannot derive an effect for the entire staff: the sample and the accounting method are not described.

In practice, such projects run into integrations, not the model. An agent needs access to the developer's internal systems—requests, documents, CRM. Each access requires permissions and logging of actions. Without this, the agent is either useless or creates risk.

It remains unknown how many employees were affected, which processes were automated, and how the time savings were measured. The implementation timeline and project cost are not stated in the source.