Tracks every close task and tells you exactly what's stuck and who owns it.
What you're looking at below is the dashboard this agent generates on every run — actual output from our demo engine, produced entirely from synthetic data. On a live engagement, it lands in your inbox before 8:00 am. Anything marked "auto-resolved" is a fix the agent can prove — a duplicate, a known-correct figure — shown so you can audit it; anything that takes judgment is flagged for a named human.