RSS
Логотип
Баннер в шапке 1
Баннер в шапке 2
Project

Severstal launches generative artificial intelligence platform

Customers: Severstal

Moscow; Metallurgical industry

Contractors: Severstal-Infokom
Product: Artificial intelligence (AI, Artificial intelligence, AI)

Project date: 2024/11  - 2025/08

Content

2025

Launch of the generative artificial intelligence platform

Severstal"" launched a platform based on generative artificial intelligence "Da Vinci." With it, company employees can create personalized assistants on the basis AI for solving work tasks and manage them in the safe contour of the company. The company announced this on September 22, 2025.

The AI assistant can generate and analyze text requests, integrate with corporate systems to quickly search for information, analyze files and extract the necessary data from them, as well as find data on the Internet. It can be used to validate or enhance content, such as reports or presentations, and to solve problems that require finding patterns in large amounts of data.

An employee can customize his assistant in an intuitive interface without having programming skills or specific knowledge. To do this, you need to create a text query, show the data source, select one of the several available large language models (LLM) and set parameters. One click - and the assistant is posted on the platform or in the corporate messenger. Other employees of the company can also use it.

Several installed models are available to platform users: for processing text and images, thinking and generating images or code. The product is the company's own development, but at the same time uses open source components, separate solutions from Russian vendors and AI models developed by other companies.

As of September 2025, 500 users are connected to the platform, they are working on 200 solutions. The plans are to provide mass access in 2026, create complete multimodality, configure hypersonalization and make a designer on the platform to create multi-agent systems in conjunction with corporate systems. Among the first results is the applied application of generative AI in the field of production technology, personnel management, in procurement, finance and repair function. For example, an assistant who processes support requests, a meeting logging system, an assistant for analyzing reports of the vibration state of the unit, and many others.

File:Aquote1.png
In 2023, we began a practical study of the possibilities of generative AI for the needs of Severstal, trying all the neural networks available at that time, both Russian and foreign. We selected the first 300 users from different features who proposed 80 hypotheses on how AI could be applied in daily work. 40 hypotheses met expectations and confirmed the effect - for example, saving time or improving content quality. Previously, the potential for the use of generative AI in real tasks of employees was limited by risks in the field of information security and the lack of the possibility of integration with corporate software. Its own platform allows you to safely use tools based on generative AI and embed them into all company processes as the basic element of the ecosystem. Our goal is to confirm the significant economic effects at the company's PnL level in 2026, put the first autonomous AI agents into productive operation and move to a large-scale implementation in all the company's processes, "commented Vladimir Sidorov, Head of the Strategy and New Technologies cluster at Severstal-Infocom.
File:Aquote2.png

Creation of an AI solution for MRO audit

Severstal-Infocom has developed a generative artificial intelligence solution for auditing maintenance and repair processes. The chatbot conducts natural language interviews on pre-configured scenarios, monitors the market, competitors and best practices, and determines the current and target level for a particular company. Severstal announced this on June 4, 2025.

The generative-based chatbot AI uses a reliability-oriented maintenance methodology (TNT) and has been strengthened data and reviewed by Severstal for more than 15 years. As of June 2025, the solution module conducts an automated analysis of current MR processes in all areas - from design documentation management to KPI repair function control, identifies suboptimal areas and generates an analytical report with a plan before achieving the strategic goals of the enterprise.

The chatbot is already used in the repair function of Severstal and allows you to minimize labor costs for regular asset screening - the complexity of assessing the current condition and tracking progress has decreased from 60 or more people's days to 1. The solution can be useful to companies of all sizes and industries. Modules can be implemented in stages depending on the maturity of processes and the needs of the enterprise.

File:Aquote1.png
Our AI assistant acts as a professional business consultant: analyzes all repair processes, identifies weaknesses and forms specific recommendations with the calculation of the economic effect. It gives enterprises a step-by-step roadmap - what organizational changes and in what sequence should be taken in order to achieve the necessary availability of equipment at minimal cost. It's not just automation - it's a decision support system to achieve the company's strategic goals. In the future, we plan to enrich it with modules for choosing the architecture of IT solutions, calculating the financial model, determining and tracking key performance indicators of the asset management system, "commented Olga Peshina, Head of Commercial IT Products at Severstal-Infocom.
File:Aquote2.png