Lukomorye IT Ecosystem Launches AI Agent Platform Oko for Business Process Automation
The Lukomorye IT ecosystem, part of Rostelecom's commercial IT cluster, has launched Oko, an AI agent platform for automating business processes. It enables the creation and management of multi-agent AI systems, integrates with corporate knowledge bases and external systems, and speeds up scenario development by 5-7 times. The platform supports various LLMs and embedding models, making it flexible and vendor-independent.
The Lukomorye IT ecosystem, part of Rostelecom, has launched the Oko platform for AI agent automation of business processes. It is designed for companies implementing AI and moving to corporate-level systems. Oko provides visual orchestration, allowing users to build workflows like a constructor, adding conditions, loops, parallel tasks, and handovers to humans. It speeds up new AI scenarios by 5-7 times compared to development from scratch, can increase operations up to three times and free up to 400 hours of team time per year, automating up to 60% of routine operations. A key feature is enabling analysts and process owners to design processes without programming. The platform includes built-in RAG mechanisms for creating and maintaining a corporate knowledge base, automatically chunking documents, creating vector indexes, and updating data. According to Denis Romanov, director of AI product development at Lukomorye, the platform is not tied to a specific language model, allowing connection of various LLMs, embedding models, and rerankers, both online and on-premises. In its class, Oko is comparable to Amazon Bedrock AgentCore and Microsoft Copilot Studio but is focused on corporate implementation in Russian companies.
- Abbreviations
- LLM = Large Language Model
- RAG = Retrieval-Augmented Generation
Source: CNews —
original
