The problem this solves
Every pipeline carries a shadow inventory of deals that stopped moving weeks ago. They inflate the forecast, absorb review time, and give reps a reason to defer the uncomfortable close-or-kill decision. The standing ritual - a manager reading down the list asking what is happening with this one - produces answers from memory, which is to say answers about what the account looked like when the deal last had energy. Nobody re-checks the account itself, because checking twenty accounts properly is an afternoon of work.
The irony is that the outside world usually holds the verdict. An account that raised money, started hiring into the buying team, or swapped leadership is a fundamentally different deal than it was at stall time, and one that quietly shrank or froze hiring is telling you something too. These signals are public and checkable. They just never get checked, because deal reviews run on CRM notes and recollection while the evidence sits outside the CRM.
A stale-deal autopsy earns its name by treating each stuck deal as a case to investigate: what changed at the account since the deal stalled, and does that change argue for one more push or a clean kill. Done for every stuck deal at once, with the verdicts posted where the team already talks, it turns pipeline hygiene from a guilt-driven chore into a standing intelligence brief.
How the mission runs
- Pull the stuck deals from Pipedrive. The agent queries Pipedrive for deals idle past your threshold - 45 days by default - capturing stage, value, owner, and last activity for each. The threshold and pipeline scope are yours to set, so the sweep can cover one team's pipeline or the whole book.
- Investigate each account's present tense. For every stuck deal, Firmographics checks the account's current size and trajectory while Hiring Data looks at open roles and team growth, particularly around the function that would use your product. The point is to compare the account as it is now against the account the deal notes remember.
- Weigh revive against kill. Each deal gets a recommendation built from the evidence: revive where the account shows fresh budget, growth, or a rebuilt buying team, kill where the signals point down or nothing has changed in months. Every recommendation cites its specific signals, so a rep can disagree with the reasoning rather than the verdict.
- Post the autopsy to #sales. The full rundown lands in Slack: one entry per deal with owner, value, days stuck, the key evidence found, and the recommendation. Formatted for a channel rather than a spreadsheet, it is readable in two minutes and tags the deal owners so the follow-up conversation starts in the thread.
The prompt
This is the exact objective the agent receives. Swap the obvious placeholders for your own domain, segment or channel and run it as-is from the console, Slack, or the API.
What comes back
A Slack post in #sales covering every deal stuck past the threshold: owner, value, days idle, what has actually changed at the account since the stall, and a revive-or-kill recommendation with its evidence cited per deal. Pipedrive itself is left untouched, so the pipeline decisions stay with the humans who own the deals while the investigation work that should inform those decisions is already done.
Make it yours
- Change the staleness threshold per pipeline stage, since 45 idle days means something different in discovery than it does in contract review.
- Schedule it for the morning of your weekly pipeline review so the meeting starts from evidence instead of memory, with the Slack thread as the agenda.
- Ask for the kill list as a separate section with suggested close-lost reasons, making the unglamorous half of the cleanup a one-click follow-through for each owner.
Frequently asked questions
Does this change anything in Pipedrive?
No. The mission reads deals and investigates accounts, and its only output is the Slack post. Closing, reviving, or reassigning deals stays entirely with the owners. If you later want the agent to apply agreed outcomes back to Pipedrive, that is a separate instruction with its own review gate.
How does the agent decide revive versus kill?
From the signals it can verify: funding events, headcount trajectory, hiring in the relevant function, and leadership changes, weighed against how long the deal has been idle. Each recommendation lists the evidence it rests on, so the team debates the facts of the account rather than the agent's opinion.
What cadence makes sense?
Weekly or biweekly, timed to land just before your pipeline review. Deals cross the staleness threshold continuously, so a standing schedule keeps the shadow inventory from rebuilding. Teams with long sales cycles often run it monthly with a higher threshold, which keeps the recommendations meaningful for their motion.