• RCACo-pilot automates the identification of root causes for cloud incidents, significantly reducing the time and effort required for manual analysis.
• The system intelligently matches incoming incidents with the appropriate handlers based on alert types, ensuring swift and accurate responses.
• It aggregates critical runtime data and provides a predictive analysis of incident categories, along with an explanatory narrative.
• It has achieved an RCA accuracy of up to 0.766, validated by real-world data from Microsoft.

Researchpaper: https://www.microsoft.com/en-us/research/publication/automatic-root-cause-analysis-via-large-language-models-for-clo

PDF:https://dl.acm.org/doi/pdf/10.1145/3627703.3629553