---
name: acct-stalled-deals
description: Reviews the stalled deals in your pipeline, with the last trace per deal, a hypothesis on the silence marked as one, a true reason to write or the decision to close the file, drafts for a small run and the pipeline updates to approve. Use for "run acct-stalled-deals", "stalled deals", "stuck deals", "deals that went quiet", "re-engage dormant opportunities", "customers who stopped replying", "clean out dead deals", "close the loop on old deals", part of the Claude for Account Executives Pack by Polar Bear.
---

# Stalled Deals Review

## When To Use
Half your pipeline has not moved in weeks and you do not dare follow up. Run this monthly, or after a pipeline review that flagged deals with no movement. It answers: for each quiet deal, do you have a true and useful reason to write, or is it time to close the file?

## When Not To Use
For one proposal sent recently, use Proposal Follow-Up. To place active deals by their evidence, use Pipeline Review.

## Inputs
- A CRM export or your notes: per deal, the stage, the date and nature of the last trace, the job role of your last contact. Salesforce in Claude (beta) reads it; any CRM works by export.
- The silence threshold you choose.
- What you know that is new about each account, with its source (your Buying Signals Brief helps).
If you have none of this, I start from the deals you name from memory with the last date you are sure of, and mark the output as a first draft.

## Approach
A review of deals with no activity, with a closing message: common practice, described generically. It uses three reasons to get in touch that serve the customer: share something useful with no ask, congratulate on a recent sourced event, suggest a conversation that is easy to decline. A quiet deal does not call for "any news?": either you have a reason that serves the customer, or the honest move is to close cleanly. The failure it prevents: a pipeline swollen with dead deals, and the same generic nudge sent to the whole list on one Monday.

## Workflow
1. I ask at most three questions: the silence threshold (there is no default), the period the export covers, and how many messages you want to write in this run.
2. Sorting: I keep the deals with no trace since your threshold and, for each, the last trace (date, what, with which job role).
3. For each deal, a hypothesis on the silence, marked as a hypothesis (project postponed, contact moved on, another supplier chosen, budget frozen), and what has changed since, with its source or "unknown". Never a hypothesis written as fact, never a judgment of the person.
4. Two options per deal, for you to decide: write, with one of the three reasons resting on a true fact, or close, with a polite message that leaves the door open. With no true reason I propose closing.
5. Drafts, for the small run you chose only: one short message per deal, to rewrite in your own words, with no "just checking in" and no reproach about the silence.
6. The pipeline update, proposed deal by deal: stage, dated next step, or closed with the known reason and "reason not given" where the customer never said. With Salesforce in Claude (beta) each write waits for your approval; otherwise you get a copy block.

## Output Format
```markdown
# Stalled Deals Review
**Silence threshold:** [set by you] · **Source:** [CRM export of date / your notes] · **Run:** [number of messages chosen]
## Quiet deals
| Deal | Stage | Last trace (date, what) | Hypothesis on the silence | Known change (source) | Your call |
|---|---|---|---|---|---|
| [account] | [stage] | [date, exchange] | [HYPOTHESIS] | [fact and source, or "unknown"] | [write / close] |
## Messages in this run
### [Deal]
Reason: [share / congratulate / suggest a conversation] · Fact behind it: [source]
[Short draft, to rewrite]
## Proposed pipeline updates
| Deal | Field | Current value | Proposed value | Approved by you |
|---|---|---|---|---|
| [account] | [stage / next step / closed] | [value] | [value] | [yes / no] |
## Decision
[Name] decides which deals to write to and which to close, and sends the messages by hand by [date].
```

## Done When
- Every deal has its last dated trace and a hypothesis marked as one.
- Every message rests on a true, sourced fact, or the deal moves to closing.
- Drafts cover the small run you chose, not the whole list.
- Every proposed pipeline change waits for your approval.

## Quality Bar
- A hypothesis about the customer is never written as a fact.
- No invented reason: a compliment or a piece of news with no source is removed.
- One message per deal, written for that deal: no variant to copy in bulk. Rules on commercial email differ by country: check with a qualified adviser.
- Closing a deal is an honest decision, not a failure to hide.
- You decide to write or to close, and you send it yourself; nothing is sent in bulk and no signal is invented.

## Next
Run acct-weekly-sales-plan (Weekly Sales Plan) to give the calls and messages you chose a place in the week.

## About the makers

This pack is made by Polar Bear, a consultancy built by ex-McKinsey founders with a dream to make AI work for People, not instead of them. We help our clients build people systems and AI-first ways of working, and we run our own company on Claude. If your team has outgrown the self-serve version, message Pauline (linkedin.com/in/paulinebertry).
