AI-Powered Incident Triage

QA-Talk
1

At our recent Ambassador Meeting, Siva Kopparapu, Senior Director, Technology at PartsSource Inc., shared how her team reduced incident triage time from approximately 60 minutes to less than 2 minutes using a custom AI agent.

Built on Claude and connected through PractiTest’s MCP, the agent automatically gathers incident-related information from PractiTest and 4 other tools, then generates a concise triage summary with the key details needed to begin investigation.

The result: faster triage, more complete information, and quicker decision-making during incident response.

image
image1920×1356 213 KB

Have questions about the implementation, integrations, or lessons learned?

Post them below—Siva will join the discussion and answer questions in the comments :speech_balloon:

3 Likes
2

cool; is it possible to share the full tools stack that was used?

3

Thanks for sharing! How did you overcome the concern about exposing data by using AI?

4

That’s awesome! I’m wondering how do you measure the quality of the AI-generated triage summaries, and how often do engineers need to correct them?

1 Like
5

Beyond the dramatic reduction in triage time, were there any unexpected lessons or limitations you discovered after deploying the agent in production?

1 Like
6

I connected Jira,Practitest,Microsoft 365,DataDog,FullStory,Github,DB2 through MCPs

1 Like