What you'll get
- 1 copy-ready blocks
- 4 min read
- The stalled-account review
How to use it
once a week, paste your account tracker into an AI assistant and ask one cold question: where is this list stalled, and where does it depend on one person? The model reads every row without favourites, and the output is a work queue, not a report.
The hack: once a week, paste your account tracker into an AI assistant and ask one cold question: where is this list stalled, and where does it depend on one person? The model reads every row without favourites, and the output is a work queue, not a report.
The prompt
The stalled-account review
Below is my account tracker. Columns: account, tier, status, roles
mapped, people in conversation (30 days), last touch, next step,
next step date.
Today's date: [date]
Review every row. Flag an account as:
- STALLED if it has no dated next step, or no touch in 30 days
(ignore accounts marked parked)
- SINGLE-THREADED if it is at replied or beyond with only one person
in conversation
- AGEING if the next step date has passed
For each flagged account, give:
1. The flag and the evidence from its row
2. The one question I should be able to answer about it and probably cannot
3. The obvious next move: a dated step, a named second role to approach,
or an honest park
Order by tier, Tier 1 first. Do not comment on unflagged accounts.
Do not soften anything.What comes back (illustrative)
"Calder Cold Chain, Tier 1: STALLED and SINGLE-THREADED. Last touch 34 days ago, one person in conversation against three roles mapped. The question you cannot answer: does the finance lead know this exists? Obvious move: the unworked operations director, this week, with a reason of her own." A dozen rows get the same treatment in under a minute, and the uncomfortable ones rise to the top, which is the point.
What to do with it
- Work the flags in tier order. Every flagged account leaves the review with exactly one of three outcomes: a dated next step, a second person to approach or a park with a revisit date.
- Use the second-stakeholder route for single-threading. The multithreading templates are built for exactly this move.
- Fix the tracker where the flag found a data hole, because many of the flags in week one are missing dates, not missing work. That is also worth knowing.
- Keep score across weeks: the flag count falling is the quiet KPI of list discipline.
The tracker format this assumes is the account penetration tracker sheet; the rules it applies are the sheet's own stall rules, applied without mercy.
Check before you trust it
- Every flag cites evidence from the row. A flag with no evidence is the model drifting; tighten the paste.
- The row count it reviewed matches the rows you pasted. Paste in batches if the list is long.
- Company names are ones you are comfortable putting into the tool under your firm's rules. If not, replace names with codes before pasting; the review works the same.
- The "obvious next move" is treated as a draft. The machine knows the rows; it does not know that the second stakeholder just changed jobs.
Three mistakes to avoid
- Running it daily. Daily flags are noise and teach you to ignore them. Weekly, before the pipeline review, is the dose.
- Treating the output as the review. The output is the agenda. The decisions, especially a park, are judgement calls the tracker cannot make.
- Letting flagged accounts skip the three outcomes. A flag acknowledged but not actioned is the tracker's version of a snoozed alarm; next week it is the same row, one week older.
Make it a habit: same morning as the weekly dashboard, flags first, moving accounts second. A quarter of this and the stalled list stops being a surprise, because nothing on it is ever more than seven days old.
Why this works on named accounts
Common questions
How often should I run an AI account review?
Weekly, before the pipeline review. Daily flags become noise and teach you to ignore them.
What counts as stalled or single-threaded?
Stalled means no dated next step or no touch in 30 days. Single-threaded means an account at replied or beyond with only one person in conversation. Change the thresholds in the prompt if your sales cycle is longer.
Can I paste client names into the tool?
Only if your firm's rules allow it. If they do not, replace names with codes before pasting. The review works the same.
More from the Academy
Account Penetration Tracker (Sheets)
For teams working a named account list.
You leave with a paste-in tracker: one row per account, six statuses, the five KPIs your weekly review needs and a rule that flags any account going stale.
Multithreading Email Templates for a Second Stakeholder
For sellers with one contact in a large account and nobody else.
You leave with a who-next card, 6 email templates, a thread log and 2 emails written out.
Want this in your inbox?
Optional. We use your email only to send you this, nothing else.