Astra Group Unveils Trusted AI Environment for the Enterprise Sector
The new technology solution can be deployed securely even at critical information infrastructure facilities.
Russian companies that operate critical information infrastructure have been slow to deploy artificial intelligence across their operations. Security requirements are the main reason for that. Most such organizations have a blanket ban on using public generative models. Specialized trusted systems could offer an alternative.
Astra Group develops systems of this kind. The company recently introduced a full suite of solutions called Astra AI. In essence, they combine infrastructure for running large language models, AI-agent development tools, a secure execution environment, and a methodology for integrating artificial intelligence into enterprise processes.
Astra Group experts argue that organizations should prioritize local solutions in which data stays inside the organization, access is controlled, and AI-agent actions are logged and constrained by security policies.
Human Control Remains
One of the company’s key products is Astra AI Code. It is an intelligent agent designed for programming. It can edit code and run build tests on its own. It does not, however, operate fully autonomously – people still make all key decisions.
“Trusted AI that is local, secure and embedded in sovereign infrastructure is the next stage of digitalization for Russian companies and the economy as a whole. Astra Group is building the full stack for it, from code to the operating system,” says Anton Shmakov, Chief Technology Officer of Astra Group.
Notably, the group first tests all of its proposed solutions on its own platforms. Using advanced technologies has significantly reduced developers’ workload. In some cases, the reduction reached 50%. At the same time, productivity can increase by 70% to 90%.
Payback in Six Months
Astra Group experts prepared illustrative financial estimates. For a hypothetical organization with a 100-person workforce, implementing trusted solutions would cost about 63 million rubles a year (about $750,000). At the same time, it would free up resources worth about 103 million rubles (about $1.2 million) that could be redirected to other tasks. The estimated payback period is about six months.
For Russia’s IT industry, the emergence of such solutions creates new opportunities to deploy AI securely while delivering measurable economic results. As of 2025, 97% of enterprises were using AI. Yet only 10% to 15% had progressed to putting models into production.
For Russia’s domestic market, the most promising areas for AI adoption are industry, government organizations, financial services, health care, and large enterprises with extensive internal documents and data.
Security First
Russia’s active development of domestic AI tools began in 2022, after Western developers left the market. One of the most significant events for the IT industry was the launch of the GigaChat large language model. It initially served as a general-purpose AI assistant, but over time it has increasingly been used for enterprise tasks. Throughout the model’s development, information security has remained a major focus.
In 2025, Russian industrial companies began actively building secure AI infrastructures. Severstal, for example, introduced a generative AI-based enterprise platform for creating personalized AI assistants for employees inside a secure environment. Early successful deployments included AI assistants for support services, automated meeting transcription, technical report analysis, as well as production operations, HR, procurement and finance.
Astra Group’s developments clearly address the market’s current needs. Rather than building individual products, the company is creating an entire ecosystem. That approach could allow Russian companies not only to compete, but also to shape their own industry trends.
