Your roadmap says a feature is "in progress." The PM is tracking it. The customer is waiting. Engineering status updates say it is on schedule. Then you check the commit history: zero commits in 5 days. This is the moment most PMs realize they have no idea what is actually happening in engineering. The problem is structural. The roadmap lives in one system and engineering work lives in another, and the PM updates the roadmap based on what engineering told them. So you have two systems of record that are frequently out of sync.
Cross-reference four systems daily
The agent runs every weekday at 9 AM and monitors four systems. It scans your roadmap (Jira, Linear, Asana) for items marked "in progress," items with an ETA this week or next, items blocked more than 7 days, and items completed in the last 7 days for velocity. For each roadmap item it finds the engineering ticket and checks status, last activity date, and open blockers. It checks GitHub or GitLab for recent commits, PR status, commit frequency, and stale branches not touched in 3+ weeks. And it scans PM communication in Slack and Notion for at-risk flags, ETA adjustments, and scope changes.
Then it asks simple, hard questions. Roadmap says in progress but no commits in 5 days, why? Engineering ticket created 6 weeks ago and still in progress, blocker or abandoned? Roadmap item with no engineering ticket at all, should there be one? Feature was supposed to be done, how accurate are ETAs really?
The six-section report
Every weekday you get a report with six sections: discrepancy alerts (the "in progress" items with no commits), stale tickets (no activity in 14+ days, critical past 21), missing engineering tickets, ETA accuracy and velocity trends, PM ownership and engagement, and a weekly summary with immediate actions.
The ETA section is quietly the most useful. It calculates how accurate your past ETAs actually were, not how optimistic they sounded: on-time percentage, average variance in days, and a 4-week velocity trend. If your team has been running at plus-1.8-days average variance, you should be baking that buffer into every current commitment.
Why the daily cross-check pays off
Here is what happened at a company I advised. A feature was scheduled to launch mid-quarter. The PM tracked it, the customer knew, the CEO mentioned it in a board update. But engineering had de-prioritized it for technical debt, and the PM did not find out until a week before launch. Then came the panic: slip the customer feature or cut the debt work, and the customer was upset either way. With a daily cross-check the PM would have seen the divergence a month earlier and raised the prioritization question with leadership then, not in a crisis.
The discrepancy alert fires within 5 days of zero commit activity, weeks before the missed deadline, when you still have options. Most days the discrepancies are minor, but 2 to 3 times a month the agent catches a major one: a stalled feature, an ETA that needs to slip, orphaned engineering work.
Connect Jira, GitHub, and Slack, run it tomorrow, and check how many of your "in progress" items actually had a commit this week.
FAQ
What data sources does it require? Jira for roadmap and engineering tickets, GitHub for commits and PRs, and Slack for PM updates. Salesforce and Gong are optional for customer context. It runs weekdays at 9 AM.
How does it prevent the quiet slip? The discrepancy alert fires within 5 days of zero commits on any 'in progress' item, giving you weeks to check blockers, de-scope, or adjust the customer commitment proactively.