МТС запустила Метан: ИИ для поиска данных в каталоге из 500 тысяч таблиц

MTS has created a system called Metan — an intelligent assistant that helps analysts find the necessary data in a corporate catalog of 500 thousand tables simply by asking a question in Russian.

A Catalog That Doesn't Answer Questions

MTS has accumulated enormous volumes of data. The company's data catalog contains over 500 thousand registered tables. Every day, hundreds of specialists work with this data: analysts, data engineers, machine learning specialists who build data marts for ML models.

But here's the problem: the catalog works well only if you already know what you're looking for. Need a table with customer information? You enter a query, the catalog gives you field descriptions, the owner, the schema.

But if you have a business question like "which tables contain information about customers in the Moscow region for the last quarter?" — the catalog won't help. An analyst has to figure it out independently: read documentation, consult with colleagues from different teams, study relationships between sources.

This can take hours or even days.

Why a Neural Network Can't Handle It Alone

It's logical to assume that a neural network could solve this task — just ask it where to find data. But there's a critical problem: AI relies only on the information it's been given. If the metadata doesn't contain explicit information about relationships between tables, about how data from one source relates to another, or about what each term actually means in the company's context, even the most advanced model will get stuck.

At MTS, however, there's a rare coincidence of two circumstances. Over 15+ years of systematic work on data governance, the company has developed serious expertise in how to properly describe metadata, structure relationships, and document business processes in data. And at the same time, LLM models have appeared that can work with such structured semantic layers.

Metan System: Combining Data Governance and AI

Metan is a name that abbreviates the words "metadata" and "analytics." It's a pilot system that MTS is testing right now. The system acts as an intelligent interface to corporate data: you ask a question in human language, the system understands which tables and sources you need, and provides an answer. This works precisely because metadata at MTS isn't just a set of descriptions. It's a full knowledge graph:

• Tables linked to their descriptions and field structure
• A glossary of the company's business terms
• Relationships between tables and data sources
• Information about responsible parties and owners
• History of dependencies and relationships between data marts

On top of this graph is a semantic layer that's understandable to the LLM. The model can "see" not just words and descriptions, but meanings and relationships between them.

What This Means for Business

Metan is not the first attempt to automate data discovery in large companies. But it's one of the first systems that actually works because it's built not on pure AI, but on a combination of expensive data governance work and modern LLM capabilities. For companies like MTS, this means accelerating analyst work — less time searching for sources and coordinating with colleagues, more time on actual analysis. For the entire industry, this is a signal: agents work better not with chaotic data, but with well-organized and well-described data. Order and AI are not enemies, but partners.