MOPZA26 Online · October 21–22, 2026

Who Changed That?

An AI Watchdog for Risky Changes Across Your Stack

An AI watchdog that reads your audit logs, flags risky changes, and alerts your team before the damage spreads.

Make it your own

Build the watchdog

Flows, system prompt, scoring rubric, and lessons learned. Everything you need to get started.

Read the build notes

Take it with you

The deck from Mopza26, with the architecture, examples, and takeaways from the session.

Watch on Mopza26 Attendee access required

01 / Read
Read the audit log
Pick up the changes your platform already records.
02 / Assess
Assess the risk
Use a reasoning layer and a rubric to flag what matters.
03 / Alert
Alert the team
Send risky changes to the channel your team already uses.

A separate volume monitor catches spikes the AI might miss.

A four-second CRM change triggered 8,400 bad MQL notifications. Finding the cause took most of a day. This is the system I built to close that gap.