At the international conference AI Journey, Sber presented the next stage in the development of its Process Mining platform. The new intelligent AI agent can be tested on an interactive stand using the example of the "Analysis of the tender procurement process" case.
The AI agent independently conducts an in-depth analysis of the business process. The user uploads a data set and formulates a task in natural language, for example: "Analyze the company's tendering activities." Next, the agent autonomously performs a full cycle of analytics: calculates key metrics, puts forward and tests hypotheses, identifying the root causes of inefficiencies. The result of the work will be a detailed report that contains the identified problems, their possible causes and recommendations for eliminating inefficiencies.
As explained in Sber, the key value of the technology is in the multi-variant nature of the research. The AI agent simultaneously tests dozens of hypotheses and issues a complete report that helps in making management decisions. Sber noted:
At the same time, the agent does not replace the specialist, but becomes his smart assistant. He takes on routine work: establishing cause-and-effect relationships, searching for hypotheses and generating recommendations. What would take an analyst several days, the agent does in tens of minutes. The task of a person is to correctly interpret the conclusions of AI, verify them and make a final decision.