An intelligent web service consultant has been created that uses OpenAI LLM models (GPT-4.1-mini) and the RAG approach (search + generation) to accurately and contextually answer user questions.

Key benefits for business:

• High speed and accuracy of responses – due to fast searching of the internal knowledge base (via FAISS) and additional enrichment via LLM.
• Integration with external sources – an MCP agent has been implemented for automatic search of additional information on the internet if local data is insufficient.
• Flexible agent logic is built using LangGraph, allowing for easy adaptation of behavior to any Business processes.
• Observability and analytics via Langfuse — every step of the model’s operation is tracked, providing transparency and quality control.

The service is easily expanded to suit various niches (medicine, law, education, tourism, consulting, and other services) and can be deployed both on the company’s server and in the cloud.