We are still trying to address the issue of upstream LLM latency when creating subjects during automation runs, which will cause the automation to fallback to its original title. We already introduced new resiliency this morning, but we want to keep this status alive for the time being.