Mean Time to Recovery (MTTR) and Mean Time to Detect (MTTD)—incident response metrics that AI tooling can improve (detection, diagnosis) or worsen (novel failure modes).
MTTR measures how fast you restore service after an incident; MTTD measures how fast you notice. AI observability (Datadog LLM Observability, Bits AI) can reduce both for AI-powered systems.