GTM AI How to Guide: Pipeline Intelligence Engine
Turn your CRM export into a complete pipeline forecast and recommendations — in one prompt.
Turn your CRM export into a complete pipeline forecast and recommendations — in one prompt.
Play created by Jonathan Kvarfordt, Head of Marketing at Momentum.io. You can find more of his GTM AI plays at gtmaipodcast.com.
FOR: RevOps & Revenue Leaders | TIME SAVED: 6–8 Hours Per Week | OUTPUTS: 3 Deliverables
01 — What It Does
Most revenue teams are still manually producing weekly pipeline forecasts — pulling from Salesforce, formatting spreadsheets, copying numbers into slides. The Pipeline Intelligence Engine play using Claude, replaces that process entirely. You give it your CRM data. It gives you everything you need for the forecast meeting.
The Claude skill runs three parallel analyses on your pipeline data and produces three files in one prompt:
Deliverable:
What’s Inside Pipeline Health Spreadsheet (.xlsx) Stage conversion rates, velocity metrics, deals at risk, week-over-week movement, and per-rep breakdown — all with auditable formulas.
Interactive Dashboard (.html) Funnel visualization, pipeline by rep vs. quota, coverage ratio gauge, and a sortable deals-at-risk table. Renders in any modern browser.
Coaching Priorities Document (.md/.docx) The 5 deals most likely to slip, rep performance highlights, and specific talk tracks for 1:1s — based on your actual data, not templates.
02 — Prerequisites
Before running the skill, confirm you have the following:
CRM data access — A CSV or spreadsheet export from Salesforce, HubSpot, or any CRM. Alternatively, connect via CRM connector or MCP integration.
Pipeline export — Include active opportunities with: deal name, stage, amount, close date, owner, created date, and last activity date. The more fields, the richer the output. If you have your CRM connected, it can pull directly from your CRM. If not, you will want to include this in your export.
Quota context — Your current quarter’s quota target and rep roster. Add these as a reference file in your project folder — the skill gets significantly smarter with this context.
Skill installed — The Pipeline Intelligence Engine skill folder must be uploaded to Claude Cowork. See Setup below. You can find the skill here.
Watch out: If your pipeline export is missing close dates or rep assignments, the skill will flag the gaps and work with what’s available — but stage velocity and rep analysis will be limited. Clean data going in means better coaching insights coming out.
03 — Setup
Set up takes about five minutes. You only need to do this once.
Install the Skill
Open Claude Cowork and go to Settings > Capabilities > Add Skill.
Upload the skill folder — the Play-01-Pipeline-Intelligence-Engine directory.
Confirm it appears in your active skills list.
Prepare Your Project Folder
Create a dedicated folder for your pipeline work. Place the following files inside:
Your pipeline data export (CSV or .xlsx from your CRM)
Your quota sheet or forecast template — optional but recommended
A rep roster with names and quota targets — optional but recommended
Connect Your CRM (Optional)
If you prefer live data pulls over manual exports, connect Salesforce or HubSpot via the built-in connector or MCP integration in Cowork settings. Once connected, the skill can pull pipeline data directly — no export required.
04 — How to Use
Once the skill is installed and your data is ready, the process runs in five steps.
Step 1 — Point Cowork at Your Folder
Select your pipeline project folder in Claude Cowork. This tells the skill where to read data from and where to save the three output files.
Step 2 — Trigger the Skill
Use any of these prompts to start the analysis:
“Run the Pipeline Intelligence Engine on my pipeline data”
“Analyze my pipeline”
“Create a pipeline health spreadsheet, interactive dashboard, and coaching priorities doc from this export”
Step 3 — Answer the Setup Questions
The skill will ask a short set of questions on the first run. Have these ready:
What CRM or data source are you using?
What is your current fiscal quarter and quota target?
Do you want analysis by rep, by team, or both?
What is your average sales cycle length?
Any specific deals or segments you want highlighted?
You can skip any questions you have already covered in your prompt or project folder context.
Step 4 — Review Your Three Deliverables
The skill generates all three files in parallel and saves them to your project folder with today’s date in the filename:
pipeline-health-[date].xlsxpipeline-dashboard-[date].htmlcoaching-priorities-[date].mdor.docx
Review each file. If any section needs refinement — different chart format, additional rep breakdowns, more specific coaching talk tracks — ask for edits in natural language. The skill will update the relevant file.
Step 5 — Make It Recurring
Once you are satisfied with the output, say:
“Save this as a weekly workflow” or “Package this as a plugin my whole RevOps team can use”
The skill will configure the workflow to run automatically — typically every Monday morning. Your team can trigger it the same way without needing to re-enter context each week.
05 — Example Use Cases
Weekly Pipeline Review
Export your Salesforce pipeline on Monday morning, drop it in the folder, and trigger the skill. In 5 minutes, you have your health spreadsheet for the forecast call, an interactive dashboard for the CRO, and coaching talking points for every manager’s 1:1. No manual formatting, no Vlookups.
Sales Leader Forecast Prep
Before your weekly forecast meeting, run the skill on the latest pipeline pull. The coaching priorities doc tells you exactly which deals to probe and what questions to ask each rep. You walk in knowing where the risk is — not finding out in the meeting.
Board-Ready Pipeline Metrics
The board wants pipeline health data. Run the skill and hand them the dashboard — funnel conversion, coverage ratios, and velocity from your actual data, not a static slide from last month.
New RevOps Hire Onboarding
Share this skill with a new RevOps team member. They generate the same reports the senior team produces on day one. No training required on which spreadsheets to pull or how to calculate pipeline velocity.
Multi-Segment Analysis
Run it on enterprise pipeline one day, mid-market the next. Compare the dashboards side by side to see where conversion rates differ and which segment needs different coaching.
06 — Best Practices
Add Context to Your Folder
The richer your input, the richer your output. Before triggering the skill, drop in your quota sheet, rep roster, and last quarter’s pipeline data for comparison. A 10-minute setup saves you an hour of back-and-forth on the outputs.
Verify Formulas on First Run
All calculations in the pipeline health spreadsheet include auditable formulas. On your first run, spot-check two or three numbers against your CRM directly. Once you trust the math, you can move faster every week.
Iterate in Natural Language
If a chart is not quite right, or you want the deals-at-risk table sorted differently, just say so. The skill responds to plain-English requests: “Sort the deals at risk by amount descending” or “Add a tab that breaks out pipeline by close month.”
Set Up the Recurring Workflow
The skill saves 6–8 hours per week — but only if you actually run it every week. Setting up the recurring workflow removes the activation energy. Schedule it for Monday morning, before your first meeting. The outputs will be waiting for you.
Handle Edge Cases Proactively
If you are working with fewer than 20 deals, the skill shifts automatically from statistical analysis to deal-by-deal review. For multi-segment pipelines, decide upfront whether you want combined or separate analysis — combined is faster, separate is more precise for coaching.
Watch out: Never use placeholder or synthetic data with this skill. It is designed to work from your actual pipeline — and will flag if it detects anything that looks like demo data. Coaching insights that don’t come from real deals aren’t useful coaching insights.
Master Checklist
Setup
Install the skill — Upload the Pipeline Intelligence Engine folder to Claude Cowork via Settings > Capabilities > Add Skill.
Prepare your folder — Add your CRM export, quota sheet, and rep roster.
Connect your CRM — Optional — link Salesforce or HubSpot for live data pulls.
Each Run
Export pipeline data — Pull the latest from your CRM, or let the connector do it.
Trigger the skill — Say “Analyze my pipeline” or “Run the Pipeline Intelligence Engine.”
Answer setup questions — CRM source, quota target, analysis scope, sales cycle length.
Review three deliverables — Spreadsheet, dashboard, coaching doc — one pass before the forecast meeting.
Iterate if needed — Ask for adjustments in plain language.
Ongoing
Save as recurring workflow — Set it to run every Monday morning.
Spot-check formulas — Verify calculations against source data on first run of each quarter.
Read the coaching doc before 1:1s — Not during — give yourself time to internalize the talking points.
 |
|