How to Use Claude AI for Pipeline Review and Deal Prioritization

Section 01

Your pipeline review is broken — here's why deals keep slipping.

Every sales team runs pipeline reviews, and most still do it the same painful way. A manager exports CRM data into a spreadsheet, spends an hour trying to build dynamic reports just to see what changed since last week, then scrolls through 40 deals asking reps for updates. The notes from that meeting? Nobody opens them again until the next one.

Here's the catch: deals that should have been flagged two weeks ago are still sitting quietly in "Negotiation." The ones that need executive attention get buried under a list of smaller opportunities. And the forecast ends up being a guess wearing a number.

The core problem: pipeline reviews are subjective. Two managers looking at the same data will prioritize different deals. Claude applies the same criteria to every deal, so the review doesn't depend on who's running it.

Common

for a slipping deal to go unnoticed for weeks

Hours

typically spent scrubbing a pipeline spreadsheet by hand

Minutes

once Claude has structured the review for you

What Claude does differently

Prioritizes by risk

Claude ranks deals by staleness, missing next steps, and stage-to-close ratio — not just by value.

Flags hidden risks

Deals with no activity in 14+ days, deals stuck at the same stage, deals with no identified decision-maker.

Generates action items

For each flagged deal, Claude suggests the specific next step — not generic advice, but context-driven actions.

Structures the forecast

Claude sorts deals into Commit, Best Case, and Upside based on the signals in your data — not gut feeling.

Section 02

The workflow: from raw pipeline data to a structured review.

This workflow takes your CRM export or pipeline spreadsheet and turns it into a prioritized review with risk flags and action items. No custom tools needed — just Claude and your data.

Step 1: Prepare your data

Export your raw pipeline data from your CRM, or copy it from your current forecasting spreadsheet. Merging data from multiple regional CRMs? You might need an XLOOKUP first, to align account owners and target values before Claude can work with them cleanly. Claude works best when it gets these specific columns:

Column	What it contains	Why Claude needs it
Deal name	Account or opportunity name	Identifies the deal
Stage	Current pipeline stage	Detects stalled deals
Value	Deal size in currency	Prioritizes by impact
Close date	Expected close date	Flags overdue deals
Last activity	Date of last meaningful contact	Identifies cold deals
Next step	Planned next action	Flags deals with no plan
Owner	Rep assigned to the deal	Groups by rep for coaching

Anonymize first: replace real company names with Company A, Company B, and so on. Replace contact names with Contact 1, Contact 2. Keep the numbers and dates real — those are what Claude actually needs to analyze.

Step 2: Paste into Claude with the prompt

Copy your pipeline data and paste it directly after the prompt. Claude reads tables, comma-separated values, or plain lists. The prompt tells Claude exactly what to analyze and how to structure the output.

Step 3: Review and act

Claude returns a structured review. You scan the flagged deals, check the risk assessment against what you already know about each account, and decide on actions. What used to eat an afternoon now takes a coffee break.

Without Claude

You scroll through 40 deals in a spreadsheet, trying to remember which ones need attention. You spend hours on it and still miss the deal that's about to slip, because the last activity was three weeks ago and nobody noticed.

With Claude

Claude scans every deal, flags the ones that need immediate attention, explains why each one is at risk, and suggests the next step for each. You spend your time on decisions, not on data gathering.

Section 03

The prompt: copy, paste your data, and get a structured pipeline review.

Pipeline review prompt

You are a senior sales operations analyst. I am going to paste my pipeline data below. Analyze it and produce a structured pipeline review.

For each deal, assess:
1. Risk level (High / Medium / Low) based on: days since last activity, whether close date has passed, whether a next step is defined, and stage progression
2. Specific reason for the risk rating
3. Recommended next action

Then organize your output as:

DEALS NEEDING IMMEDIATE ATTENTION
(High risk deals — list each with: deal name, value, risk reason, recommended action)

DEALS ON TRACK
(Low risk deals — brief summary only)

FORECAST SUMMARY
Categorize all deals into:
- Commit (high confidence, close this period)
- Best Case (likely but not certain)
- Upside (possible but uncertain)
Show total value for each category.

FLAG any deal where:
- Last activity is older than 14 days
- Close date has passed
- No next step is defined
- Deal has been in the same stage for 30+ days

Do not invent information. If data is missing for a deal, flag it as "Incomplete data — needs rep input."

Here is my pipeline data:
[PASTE YOUR PIPELINE DATA HERE]

What to expect from the output

Immediate attention

Deals flagged as high risk with specific reasons — "no activity for 21 days", "close date passed 2 weeks ago", "no next step defined."

On track

Deals with recent activity, defined next steps, and realistic close dates. Brief summary so you can skip quickly.

Forecast

Deals sorted into Commit, Best Case, and Upside with total values. Paste this structured output into your sheet and use IF/IFS logic to trigger alerts for high-risk accounts.

Data gaps

Deals where information is missing — "no close date", "no owner assigned." These need rep input before you can assess them.

Review checklist

After Claude returns the review:
1. Scan the high-risk deals first — do you agree with the flags?
2. Check the forecast categories — does Commit feel realistic?
3. Assign action items to reps for each flagged deal
4. Note any deals where Claude flagged missing data
5. Use the structured output in your team review meeting

Section 04

Quiz: test your knowledge.

1. You're prepping for the weekly forecast call and want every deal judged by the same criteria, not just the ones you happen to remember. What does Claude actually give you here?

A structured, consistent read on every deal, so nothing gets judged from memory alone

It replaces the sales manager's judgment entirely

An automatic update to your CRM

2. You're about to paste your pipeline spreadsheet into Claude, and it still has real client and contact names in it. What do you do first?

Paste it as-is — Claude only needs the numbers anyway

Strip out all the numbers before pasting

Replace company and contact names with placeholders, and keep the numbers and dates

3. A deal has sat in "Negotiation" for six weeks, with no logged activity and a close date that's already passed. How would Claude flag it?

Low risk — negotiation stage means it's close to signing

High risk — stale activity, a missed close date, and a stuck stage are exactly what it looks for

It wouldn't flag it, since deal value is what matters most

4. One deal in your export is missing a close date and an owner. What does Claude do with it?

Invents a plausible close date so the forecast still adds up

Drops the deal from the review entirely

Flags it as incomplete and asks for rep input, rather than guessing

Correct answers