| The name of the base system (platform): | Artificial intelligence (AI, Artificial intelligence, AI) |
| Developers: | T-Technologies (formerly TKS Holding) |
| Last Release Date: | 2025/09/12 |
| Branches: | Information Technology |
| Technology: | RPA - Robotic Process Automation, Speech, Application Development Tools |
The main articles are:
- Virtual Assistants
- Speech Recognition (Technology, Market)
- Speech technology: On the path from recognition to understanding
- RPA (Robotic process automation)
- AI Agents (Agent AI)
2025: Launch of AI Agent for Developers
T-Technologies have created an agency mode within their own ecosystem of AI assistants, which covers the entire development life cycle - from collecting requirements to testing and supporting the finished solution. On September 12, 2025, TAdviser was informed about this in T-Bank.
In client mode, the AI assistant not only prompts parts of the code, but performs a sequence of actions on the task of the engineer. Agent mode is Qwen3-Coder-480b-based and supports up to 262,000 context tokens.
According to T-Technologies, unlike classic code prompts, agent mode turns the tool into a full-fledged "digital colleague": he understands the whole task, is able to execute entire scripts, work with the project context, taking into account the repository structure, can use utilities, commands, create files and not only. Integration into VS Code simplifies work: the AI agent is available directly in the editor, executes instructions in Russian or any other language.
The key features of the tool are integration with existing infrastructure and understanding the context of large ecosystem projects. According to T-Technologies, the agency regime accelerates development and review, combining dozens of actions into one request. It helps to perform standard operations faster, simplifies the start for beginners and increases the productivity of experienced developers, leaving more attention to the architecture and product. According to internal measurements of T-Technologies, the time for repeated operations is reduced significantly: for example, the launch of simple test versions of products takes several minutes instead of several days.
As of September 12, 2025, 80% of engineers regularly use AI tools for development within the T-Technology ecosystem, or more than 10 thousand users per month.
| "AI agents for developers have already become a global trend: such solutions have already been launched by Google, Claude Sonet, Cursor and a number of other companies. We became the first company in Russia to massively introduce its own AI agent for internal teams. This allows you to significantly speed up the launch of new products on the market and reduce the cost of creating them. According to our estimates, the time from the idea to the launch of a product in a large business is reduced to 40%, and resource savings for implementation reach at least 20%, "said Igor Maslov, Vice President of T-Bank, Director of Development of Basic Technologies and Engineering Practices. |
The ecosystem of T-Technology AI tools for development for September 2025 includes a full range of AI tools for analysis, development, testing, implementation and subsequent product support:
- AI-search: helps IT teams quickly find information: ways to fix a bug, business rules, architectural solutions.
- AI-Data: AI assistant to speed up analytical tasks. Helps you write and edit SQL queries, generates relevant scripts for ETL, and answers analytical questions directly in the context of working with data.
- Autodesc-Generates relevant descriptions of tables and fields and publishes them to the data catalog.
- T-Cover Agent: The LLM agent generates correct unit tests, increases test coverage, and protects logic from regressions.
- T-Weaver: analyzes the documentation and generates a complete set of test scenarios with structure, steps, expected results and the desired attachments.
- T-Code Review: AI code review automation assistant, covers 100% of all repositories. That's 1,000 + comments a day.
- AI Assistant: A smart AI assistant for IDE development.
- Safeliner: AI-assistant for information security.
- SRE assistant: a digital SRE engineer who himself searches for information and documents.
- Anomaly Analyzer: Detects deviations in metrics in real time and sends notifications to commands.
T-Technologies plans to bring the entire ecosystem of AI tools for developers into the B2B segment for medium and large businesses. Early registration for users will open soon. In addition, it is planned to expand the functionality of the agent mode with new scenarios, including an AI agent for data analysts.
