AI Monitoring Solution

The AI Monitoring Solution is a service that uses alerts generated by monitoring systems as a starting point and comprehensively analyzes past incident response records, operational manuals, and internal knowledge. When an incident occurs, the AI supports the initial response activities typically performed by operations personnel, including gathering information, searching for similar past cases, investigating the cause, and organizing the response strategy.

The AI organizes the alert details and presents possible causes, the scope of impact, and recommended investigation procedures and response methods. Similar past incidents and response records can be searched using RAG and incorporated into the analysis results. This enables operational know-how that previously depended on individual experience and memory to be accumulated as organizational knowledge that can be shared and utilized across the company.

Analysis results can be output as reports in a standardized format and automatically registered in ticket management systems such as Backlog. By leveraging Amazon ECS, AWS Lambda, Amazon Bedrock, and Amazon RDS for PostgreSQL, we support the optimization of the entire incident response process—from integration design with monitoring systems and knowledge-base development to the construction of an AI analysis platform and automated ticket creation.