At the end of our pipelines acceptance-tests, upgrade-tests and cluster-upgrade-tests we send a Slack notification to #tektoncd-pipeline-ci with pass/fail information for each run of test suite. This notification also contains link to the pipeline run logs for convenience.

QEs need to review logs to understand which tests failed and what might be the cause.

Report portal was supposed to help with test failure classification but we never implemented it fully.

Research the possibility of creating a task and/or agent that would use AI to analyze logs and send summary to the Slack channel.

There is already an initial implementation of Tekton MCP server that might be handy for this - https://github.com/openshift-pipelines/mcp-tekton