A weekly pipeline rhythm needs a Monday scan, a coverage check, a hard look at single deals, and a Friday summary that doesn’t guess at causes. This guide has a Claude prompt for each of those six jobs and a line on what to check before you trust the output.
The parts run Monday to Friday. The numbers are yours; the prompts only arrange them.
In this guide
- Pipeline review: a Monday scan and a plan for the week
- Am I covered? A pipeline-versus-quota check you can read as a range
- A skeptical sales manager for your own deal
- A blameless review of a deal you lost
- Weekly ops: a Friday summary and a check for process leaks
- Explain this week’s numbers in plain language, without claiming causes
Pipeline review: a Monday scan and a plan for the week
Most pipelines have deals that need you this week and deals that will never close. A Monday scan helps you tell them apart.
The weekly scan
Paste a table with the deal, stage, amount, close date and last activity. The prompt flags the three clear problems: no next step, no recent activity, and a close date in the past.
Here is my open pipeline: [PASTE TABLE WITH DEAL, STAGE, AMOUNT, CLOSE DATE, LAST ACTIVITY]. Flag deals with no next step, no activity in [14] days, or a close date in the past. List them by risk, with one sentence on why for each.
Check before you use it: Check the CRM before you chase anyone. The data may be stale, not the deal.
Plan the week from the pipeline
Then ask for the five deals most worth your time and one next action for each. The useful half of this prompt is the second list: the deals to stop spending time on, with a reason.
Given this pipeline [PASTE], pick the five deals most worth my time this week and the single next action for each. Then list the deals I should stop spending time on, with a reason for each.
Check before you use it: Claude doesn’t know your calendar or your relationships. Adjust the plan.
The catch
A stale deal and stale data look the same in a table. Check the CRM before you chase anyone. Claude also doesn’t know your calendar or your relationships, so adjust the plan to fit your week.
Related: If the CRM doesn’t enforce it, reps will skip it
Am I covered? A pipeline-versus-quota check you can read as a range
Every seller wants to know if the pipeline covers the number. The honest answer is a range, with the assumptions written down.
This prompt takes your quota, what you have closed, your open pipeline by stage and your past stage-to-close rates, or the word “unknown.” It estimates what you can expect to close, shows the math and states every assumption. The assumptions are the useful part, because they show where the estimate is weak.
My quota is [AMOUNT] for [PERIOD]. I have closed [AMOUNT]. My open pipeline by stage is [PASTE]. My past stage-to-close rates are [PASTE, OR WRITE “UNKNOWN”]. Estimate what I can expect to close and whether I am covered. Show the math and state any assumption you made.
Check before you use it: Rates from a small sample are rough. Read the result as a range, not a promise.
How to use it
- Use real numbers from your CRM. Say “unknown” where you don’t have them.
- Read the assumptions before you read the answer.
- Treat the result as a low-to-high range, then decide what to do about the low end.
The catch
Rates from a small sample are rough. If you have closed only a handful of deals, the model is working from anecdotes. Say so in the prompt and read the output with that in mind.
A skeptical sales manager for your own deal
The deals you believe in are the ones you question least. This prompt asks Claude to do the questioning.
It plays a skeptical sales manager, asks the seven hardest questions about the deal one at a time, and after you answer, says whether the deal belongs in the forecast and why. Asking one at a time matters. It makes you answer each question before the next one arrives.
Act as a skeptical sales manager reviewing this deal: [NOTES]. Ask me the seven hardest questions about it, one at a time. After I answer, tell me whether it belongs in the forecast and why.
Check before you use it: Answer honestly. The point is to find what you don’t know.
How to use it
- Paste your real notes, including the parts that make you nervous.
- Answer each question out loud or in writing before you read the next.
- Write down every question you couldn’t answer. Those are your next calls.
The catch
The point is to find what you don’t know, not to win the argument. If you catch yourself defending the deal, stop and note the question you are avoiding.
A blameless review of a deal you lost
Most loss reviews turn into a hunt for a reason or a person. A blameless review looks for the moment the deal changed direction.
This prompt takes your notes and timeline, then finds where the deal turned, which signals were missed and one change to the process. It asks Claude to quote the evidence for each point. A point without a quote is a guess.
We lost this deal. Notes and timeline: [PASTE]. Do a blameless review. Find where the deal changed direction, what signals we missed, and one thing to change in our process. Quote the evidence for each point.
Check before you use it: Keep it about the process, not people. Drop anything Claude can’t back with a quote.
How to use it
- Paste the real timeline, including emails and call notes, not a summary of it.
- Drop any point that Claude can’t back with a quote.
- Pick one process change and write it down where your team will see it.
The catch
Keep it about the process, not people. One lost deal is a single case, so treat the finding as a hypothesis and look for the same pattern in the next loss before you change how you work.
Related: Find your ideal customer in the deals you already closed
Weekly ops: a Friday summary and a check for process leaks
A week that ends without a summary starts the next one without a plan. These two prompts take about ten minutes.
The Friday summary
Paste in what happened and ask for wins, stuck deals and the first thing you will do on Monday. The 150-word limit keeps it honest.
Here is what happened this week: [PASTE ACTIVITY AND NOTES]. Write my Friday summary under 150 words: wins, stuck deals, and what I will do first on Monday.
Check before you use it: Add anything Claude couldn’t see, like a hallway conversation.
Where does the process leak?
Describe how you run a process today and ask for the five steps where deals or information are most likely to get lost, with one simple check for each.
Here is how we run [PROCESS] today: [DESCRIBE]. Find the five steps where deals or information are most likely to get lost, and suggest one simple check for each.
Check before you use it: Walk through the steps with someone who does the work.
The catch
Add what Claude couldn’t see, like a hallway conversation. And for the process review, walk through the steps with someone who does the work. They will know which of the five are real.
Explain this week’s numbers in plain language, without claiming causes
Numbers go up and down. The owner wants to know why, and the honest answer is usually “we should check.”
This prompt takes this week’s numbers against last week’s and explains what changed in plain language for a small-business owner. It forbids claiming causes and asks for possible causes as questions. That keeps a guess from turning into a story.
Here are this week’s numbers against last week’s: [PASTE]. Explain what changed in plain language for a small-business owner. Don’t claim causes. List possible causes as questions I should check.
Check before you use it: Look up the real cause. Claude is listing possibilities, not answers.
How to use it
- Paste both weeks in the same layout so the comparison is fair.
- Read the plain-language summary first, then the list of questions.
- Look up the real cause for the two biggest changes before you share the update.
The catch
Claude is listing possibilities, not answers. Don’t pass the questions to your boss or your client as findings. Check them first, then write what you found.
Where to go next
Every prompt here comes from a set of 50 I wrote for the moments in a sales week. A free sample of 10 and the full pack are in the shop.
If you’d rather have this built into your CRM than run it by hand, that’s what RevLathe does.
I’m Seth Levanen. I find the AI shortcuts that keep the quality and drop the busywork. Email me if you want to talk about it.