It’s Friday at 3:45 PM. Your forecast call is at 4:30. You have a raw pipeline export from Salesforce, 23 tabs open, and 45 minutes to turn messy CRM data into a coherent story your CRO can defend to the board.
You’re not doing analysis right now. You’re doing translation — converting fields and dollar amounts into a narrative that makes sense to someone who wasn’t in every deal review this week. That’s what RevOps does every Friday. It’s also the most eliminable part of the job, and nobody’s eliminated it yet.
That changes with Claude Cowork.
Cowork is Anthropic’s persistent agent system — Claude that runs recurring tasks, maintains context across sessions, and works across Desktop, iOS, and Android. You set up a thread once, teach it your pipeline structure, your stage definitions, your thresholds. Every week after that, you paste in your data and get a brief back. The translation is done before you’ve finished your coffee.
What You’ll Build
Four persistent threads that cover the RevOps operating cadence:
Weekly Forecast Digest — turns your CRM export into a structured forecast brief, deal by deal, with the number to defend
Pipeline Health Monitor — tracks velocity and conversion week-over-week, flags deals breaking pattern
Revenue Cadence Prep — pre-builds your weekly revenue review agenda and talking points
Exec / Board Reporting Brief — synthesizes 4 weeks of data into a board-ready narrative
Plus a fifth for power users: Territory and Quota Tracker.
Step 1: Setup
Time required: 15 minutes. No developer.
Access: Claude.ai Pro or Team plan. Desktop app recommended for RevOps work — you’ll be pasting exports frequently, and the desktop clipboard handling is faster.
Thread naming convention: For RevOps, name threads to match your reporting cadence. This isn’t aesthetic — it’s operational. You want to open the right thread without thinking.
Weekly Forecast — [Current Quarter]Pipeline Health — [Current Quarter]Revenue Cadence PrepBoard Brief — [Month]
Create a new thread for each quarter’s forecast work. Don’t carry Q2 data into Q3 threads — the context gets muddy and the baselines shift.
Step 2: The Weekly Forecast Digest Thread
The CRM export — under 2 minutes from HubSpot or Salesforce:
HubSpot: Sales > Deals > All Deals > filter by Deal Stage (exclude Closed Won/Lost) > Actions > Export. Select: Deal Name, Amount, Close Date, Deal Stage, Owner, Last Activity Date, Next Step. CSV.
Salesforce: Reports > New Report > Opportunities > filter by Stage (open only) > add columns: Account Name, Amount, Close Date, Stage, Owner, Last Activity, Next Step > Run > Export Details > CSV.
Open the thread. Paste this prompt first — you only do this once per thread:
FORECAST DIGEST SETUP
You are my weekly forecast analyst. Every week I'll paste a pipeline export from [HubSpot/Salesforce].
Your job is to produce a structured forecast brief I can use on our Friday forecast call.
My pipeline stages and their meaning:
- [Stage 1 name]: [what this means — e.g., "Discovery complete, qualified"]
- [Stage 2 name]: [what this means — e.g., "Proposal sent, verbal interest"]
- [Stage 3 name]: [what this means — e.g., "Contract out, high confidence"]
- [Stage 4 name]: [what this means — e.g., "Commit — should close this quarter"]
My forecast call format:
- "The number" = total commit + likely best case
- Deal tiers: Commit (closes this period, 90%+), Best Case (could close, 60-80%), Pipeline (below that)
- Risk flags: deals with no activity in 14+ days, close dates that slipped from last week, missing next steps
When I paste an export, produce:
1. The Number (commit total + best case total)
2. Commit list — deal name, amount, owner, close date, one-sentence status
3. Best case list — same format, with one risk flag per deal if applicable
4. Top 3 risks that could move the number down
5. Gaps — deals with no activity, missing next steps, or stalled in stageStep 3: Your First Forecast Brief
Paste your CSV export directly into the thread. Claude reads the columns and maps them to the structure you defined.
What you get back: a formatted forecast brief with the number at the top, deal-by-deal commentary, risk flags highlighted, and the gaps list. Before your call, not during it.
The first output won’t be perfect — your stage names are specific to your setup, and Claude will ask clarifying questions. Answer them in the thread. By week three, it knows your pipeline the way a sharp analyst who’s been in every deal review would.
Keep reading with a 7-day free trial
Subscribe to GTM AI Podcast & Newsletter to keep reading this post and get 7 days of free access to the full post archives.

