Abstract
Artificial intelligence (AI) troubleshooting agents can lose root cause analysis (RCA) investigation state when conversation history is compacted or an investigation is interrupted and resumed. To address the challenges of maintaining RCA investigation state, the techniques presented herein implement a memory system that maintains structured RCA memory outside the chat transcript. The memory tracks diagnostic coverage, unexplored evidence, and candidate root causes; refreshes stale state before compaction; and updates the investigation when new evidence is added. A candidate-cause ledger supports final causal adjudication, reducing the likelihood that a downstream event is mistaken for the underlying root cause.
Creative Commons License
This work is licensed under a Creative Commons Attribution 4.0 License.
Recommended Citation
Cehreli, Necati, "ROOT CAUSE INVESTIGATION MEMORY FOR ARTIFICIAL INTELLIGENCE (AI) AGENTS", Technical Disclosure Commons, ()
https://www.tdcommons.org/dpubs_series/11380