A Weekly Pipeline Review That Flags What's Slipping
Turn your pipeline export into a weekly action plan: which deals to push, which are stalling, which have hygiene problems, and the single-threaded risks before they cost you.
A pipeline review should tell you what to do this week — not just what the numbers are.
Staring at a deal list doesn't surface the deal that's quietly dying or the one riding on a single contact. Store what 'at risk' means for you, and the review turns a flat export into a ranked action plan: push these, save these, fix these.
The deal that slips isn't usually the one you worried about. It's the one that looked fine on the board while the only champion went quiet.
A pipeline review is only useful if it ends with a list of moves, not a list of deals.
The plan is a prompt for your judgment, not a verdict. You know which deals are real; the review makes sure none are quietly slipping while you focus elsewhere.
| Complexity | Easy |
| Tools needed | Claude Pro or above, Claude Desktop → Cowork mode, on macOS or Windows, Claude Projects |
| Time to build | ~20 min first time · ~5 min per week after that |
| Best for | Founder or sales lead who owns the pipeline — B2B services and software |
What this solves
Weekly pipeline reviews that end in a vague plan while the deal that's quietly dying stays invisible until it's too late to save.
The problem
Monday’s pipeline review is supposed to tell you where to spend the week. In practice you scroll the deal list, nod at the big ones, and walk away with a vague plan. Nothing on the board screams “I’m about to slip” — the deal that dies next month looks fine today, because the warning signs are subtle: no activity in two weeks, a close date that’s been pushed twice, a single contact who’s gone quiet. By the time it’s obviously at risk, it’s usually too late to save. The review felt productive but didn’t change what you’d have done anyway.
The fix is a review that does the diagnostic work for you — surfacing the subtle risks and ending with a ranked list of moves.
Ingredients
- Claude Subscriptions — Cowork and Projects aren't available on the free plan
- Platforms & Modes — Cowork runs in the desktop app only — not on web or mobile. This recipe analyzes your pipeline and builds an action plan — multi-step work that runs in Cowork
- Claude Projects — Your stages, your definition of risk, and your process are constant. A Project stores them so every weekly review flags the same things by your standard
How it works
You review pipeline every week, and your stages and risk signals don’t change. A Project stores them so every review applies the same diagnostic and you’re not re-explaining what “at risk” means.
- Open Claude Desktop and click Cowork in the mode selector across the top (Chat · Cowork · Code).
- In the left panel, find Projects and click the + button.
- Choose Start from scratch. Name the project “Pipeline Review” and let Claude set up its folder.
- You’ll know it worked when the project appears in the left panel with its own folder and an instructions field.
Your project has a folder on your computer — that’s where the rules file lives permanently. The weekly export joins it each Monday.
- Ask Claude, right in the project: “Create a file in this project’s folder:
pipeline-rules.md. Leave it empty — I’ll fill it in.” (Or create it yourself in any text editor and save it into the project folder.) - Fill it in using the description below.
- Each week, drop this week’s CRM export into the same folder as
pipeline.csv(or connect the CRM and skip the export — see Extra Credit). - Confirm Claude can see them: ask “List the files you can see in this project.” Both filenames should come back. If they don’t, see If It Doesn’t Work.
pipeline-rules.md
Your stages, what advancement requires, and your risk signals — so the review flags by your definitions, not generic ones.
Example: “Stalled = no activity 14 days. Risk: close date pushed 2+ times, single contact only, stuck in one stage over its normal cycle. Hygiene flags: close date in the past, missing next step, deal in ‘proposal’ with no amount. Single-threaded = only one contact engaged.”
pipeline.csv
Your active deals exported from the CRM — stage, amount, close date, last activity, contacts. Or connect the CRM and skip the export.
Open your Project in Claude Cowork. Give Claude the specifics for this run, ask for the main output, then follow up for any additional pieces you want. The exact wording for each prompt — with what it’s asking for and why — is in What you actually type below.
Check three things before you run your week off it:
- The plan is moves, not observations. “Acme is stalled” is an observation. “Email Acme’s economic buyer to re-engage; if no reply by Thursday, it’s slipping out of the quarter” is a move. If the output is descriptions, ask for the specific action per deal.
- The risk flags match your signals. Confirm Claude flagged deals by the rules in your file — pushed close dates, single-threading, inactivity — not a generic sense of stale. A flag you disagree with erodes trust in the rest.
- Hygiene issues are caught. Close dates in the past, missing amounts, deals with no next step — these distort your forecast. A good review surfaces them so the numbers you report are real. Make sure they’re flagged.
Before you act: does the ranking match what you know is actually live? Claude reads the data; you know which deals are real and which the CRM oversells. Your read plus the diagnostic beats either — override the ranking where you have private knowledge.
What you actually type
Name your files explicitly the first few runs, and ask Claude to show its work on anything that matters.
“Here's this week's pipeline [attach CSV or pull from CRM]. Using `pipeline-rules.md`: rank the deals I should push this week and why, flag everything stalled or at risk by our signals, list hygiene problems to fix, and call out any single-threaded deals. End with my top five moves for the week, most urgent first.”
“Draft a multi-threading email for the two single-threaded deals to reach a second stakeholder.”
A weekly review with five labeled parts: the deals to push this week with reasons, everything stalled or at risk by your signals, hygiene problems to fix, single-threaded deals to multi-thread, and a top-five move list ranked most urgent first. Each move names a deal and a specific action. If it reads as descriptions — 'Acme is stalled' with no next step — it missed; send it back for the action per deal.
- It reads the export, not the deals. A deal the CRM oversells will look healthy in the review — your private read of which deals are real still decides.
- The flags are only as sharp as `pipeline-rules.md`. Vague rules produce generic 'stale deal' warnings you'll learn to ignore.
- It surfaces the moves; it doesn't make them. The multi-threading email still has to be sent, and the quiet champion still has to be called.
If it doesn’t work
- No Cowork tab in Claude Desktop — update the app to the latest version and confirm you’re on a paid plan; Cowork isn’t on the free tier. On Windows, Cowork also needs the Virtual Machine Platform feature enabled — if the tab still won’t appear, that’s the fix.
- Claude can’t see
pipeline-rules.mdorpipeline.csv— the files aren’t in the project’s folder, or this week’s export landed somewhere else. Open the project, check which folder it points to, and move the files there. Then re-run “list the files you can see.” - The review is observations, not moves — “Acme is stalled” gives you nothing to do Monday morning. Ask: “For each flagged deal, give me the specific action and the day it’s due.” If it still describes instead of directing, your prompt dropped the “top five moves” line — put it back.
- The flags don’t match your sense of risk — either
pipeline-rules.mdis too vague to filter with, or the export is missing the columns the rules need (last activity, contact count, close-date history). Sharpen the thresholds and re-export with the fields your signals depend on.
Extra credit
Small additions that pay back the next time you run it.
- CRM connector — connect your CRM so the review pulls live and you skip the weekly export. See the Connectors guide.
- Monday auto-review — set a scheduled task to run the review every Monday at 7am so your action list is waiting before the week starts.
- Forecast handoff — feed the cleaned pipeline into the Forecast recipe so your number is built on a board you’ve already scrubbed.
“Most lost deals were visible on the board for weeks — just never flagged as the ones to save.”
What this teaches you about Claude Cowork
The recipe is one application. The principles apply to everything you’d hand to Claude.
Diagnosis is the work. Anyone can read a deal list; the value is spotting the subtle risk before it’s obvious. Storing your risk signals lets Claude do that diagnostic every week, consistently — which is what turns a review from a ritual into a plan.
Rules make the flags yours. Generic “stale deal” warnings get ignored. Flags built on your definitions of risk and hygiene are ones you act on, because they match how you actually judge a deal.
It ends in moves, and you still decide. The recipe’s discipline is producing a ranked action list, not a report. You bring the private knowledge of which deals are real; the review makes sure nothing slips while you’re looking elsewhere.
Who this is for
Founder or sales lead who owns the pipeline in B2B services and software (5–100 employees).
The pain: Deals slip weeks after the warning signs were visible on the board
The outcome: Every Monday starts with five ranked moves instead of a scroll through the deal list
Published June 26, 2026 · 0 views