VK AI Space has announced the launch of its support for multi-agent systems, allowing users to create complex business scenarios involving a team of specialized agents managed by a leading universal agent. This innovative approach enables companies to transition from isolated AI agents to coordinated systems of digital employees. Representatives from VK shared this development with industry insiders.
The multi-agent system functions like a team of experts under a manager's direction. The leading agent receives a task from the user, understands the objective, selects the appropriate agents with the required skills, and allocates the work among them. Each agent operates within its area of responsibility and can be swiftly replaced, refined, or integrated into new processes without the need to rebuild the entire workflow. This methodology accelerates the development of intricate AI solutions while ensuring that individual agents work with the full context necessary for task completion. The multi-agent system, in turn, focuses solely on the outcomes produced by these sub-agents, thereby optimizing token usage by only transmitting essential context during the agent session. VK AI Space places significant emphasis on observability, allowing users to view both the internal steps of individual agents and the interactions among agents within a single scenario.
Additionally, AI agents on the VK AI Space platform can utilize skills—sets of instructions, steps, and rules for specific tasks described in natural language without the need for programming. These skills serve as the foundation for modern agent systems, facilitating their reuse across multiple scenarios and among various agents. This not only speeds up development but also standardizes processes and reduces the costs associated with reconfiguration. Skills can be developed internally by the company or sourced from external providers, with their execution taking place within containers in an isolated environment. This guarantees safe interactions with corporate systems and allows AI to undertake real actions without jeopardizing the existing IT infrastructure.
According to a report from Apple Hills Digital, the AI platform segment is the fastest-growing area of the Russian software market, projected to grow by approximately 50% annually, reaching 46.5 billion rubles by 2030. Businesses are increasingly willing to assign more complex tasks to AI, provided they can rely on the manageability and predictability of the results. The multi-agent system addresses these concerns by distributing responsibilities: the leading agent orchestrates the work, specialized agents execute their designated tasks, skills enable rapid capability expansion, and isolated containers ensure that each action does not harm business systems. This shift signifies a move from experimental assistant usage to genuine process delegation, simplifying the diagnostics, auditing, and optimization of AI processes.
This advancement positions VK as a key player in the rapidly evolving AI market, pushing competitors to innovate and adapt their offerings to keep pace with the growing demand for reliable and efficient AI solutions.
Informational material. 18+.
