Skip to content
Claude for Operators

CRM Intelligence & Pipeline Automation

Last verified: 14 February 2026 | Applies to: Pro, Max, Team, Enterprise (requires connectors)

In 30 seconds

Section titled “In 30 seconds”

Your CRM is full of data you never have time to analyse properly. Claude connects to HubSpot or Salesforce via MCP connectors, reads your pipeline in real time, scores leads based on patterns you define, and drafts follow-ups that reflect each deal’s actual history. This turns your CRM from a glorified spreadsheet into something that actively helps you close deals.

What it does

Section titled “What it does”

Once your CRM is connected, Claude can:

  • Read your pipeline — pull deal lists, filter by stage, owner, close date, or value
  • Analyse deal health — flag stalled deals, identify at-risk opportunities, spot patterns in win/loss data
  • Score leads — apply criteria you define (company size, engagement level, industry fit) to rank prospects
  • Draft follow-ups — write personalised emails based on each deal’s notes, stage, and last activity
  • Produce pipeline reports — weekly summaries, forecast snapshots, and board-ready pipeline overviews

Claude doesn’t replace your CRM. It sits on top and makes the data inside it useful.

How to set it up

Section titled “How to set it up”

Step 1: Connect your CRM

Section titled “Step 1: Connect your CRM”

HubSpot: Go to Settings → Connectors in Claude Desktop, search for HubSpot, and authenticate via OAuth. You need at least read access to contacts, companies, and deals.

Salesforce: Salesforce requires a custom MCP connector. Go to Settings → Connectors → Add custom connector and enter the Salesforce MCP server URL. Your Salesforce admin may need to provision API access. Check with your admin or see the connectors guide for details.

Step 2: Seed your sales context

Section titled “Step 2: Seed your sales context”

Give Claude the information it needs to understand your pipeline:

Our sales process has five stages: Discovery, Demo, Proposal, Negotiation, Closed Won. Average deal cycle is 45 days. We sell B2B SaaS — our ICP is mid-market companies (50-500 employees) in professional services. Average deal size is $36,000 ARR. Our sales team: Jake (enterprise), Mei (mid-market), and Tom (inbound). We consider a deal "stalled" if there's been no activity for 14 days.

If you have workplace memory set up, this gets stored and persists across sessions.

Step 3: Test the connection

Section titled “Step 3: Test the connection”
Pull my current pipeline from HubSpot. Show all deals in Proposal or Negotiation stage, sorted by expected close date. Include deal owner, value, and days since last activity.

Claude reads your CRM and returns a structured table. If this works, the connection is live.

How operators actually use it

Section titled “How operators actually use it”

The Monday morning pipeline review

Section titled “The Monday morning pipeline review”
Give me a pipeline health check. Pull all open deals from HubSpot and tell me:
1. Total pipeline value by stage
2. Deals closing this month — are any at risk?
3. Stalled deals (no activity in 14+ days)
4. Top 5 deals by value with next steps from the last note
5. Win rate trend — how does this month compare to the last three?

Claude returns a structured briefing you can bring to your Monday sales meeting — or forward to your sales lead with a note.

Lead scoring

Section titled “Lead scoring”
Pull all leads created in the last 30 days from HubSpot. Score each one on a 1-10 scale based on:
- Company size (50-500 employees = high fit)
- Industry (professional services, consulting, or legal = high fit)
- Engagement (opened 3+ emails or attended a demo = high engagement)
- Source (referral or inbound demo request = highest value)


Rank them and tell me which 10 leads Jake and Mei should prioritise this week.

Claude pulls the leads, applies your scoring criteria, and returns a prioritised list with a one-line rationale for each score. Distribute the list to your reps or drop it into Slack.

Automated follow-up drafting

Section titled “Automated follow-up drafting”
Pull all deals in the Proposal stage where the last activity was more than 7 days ago. For each one, draft a short follow-up email from the deal owner. Reference the specific proposal we sent (pull the details from deal notes) and suggest a concrete next step. Keep each email under 100 words.

Claude drafts personalised follow-ups — not generic templates, but emails that reference each deal’s actual context. Review, adjust, and send.

Win/loss analysis

Section titled “Win/loss analysis”
Pull all deals that closed in Q4 — both won and lost. Analyse the patterns:
- Average cycle time for wins vs. losses
- Stage where most deals die
- Common objections in lost deal notes
- Which rep has the highest conversion rate from Demo to Proposal?
- Any patterns by industry or company size?


Summarise findings and give me three actionable recommendations.

This is the analysis most sales teams never get around to doing. Claude processes the data in minutes and surfaces patterns that would take hours to find manually in your CRM’s built-in reporting.

Section titled “Related”

Something wrong or outdated? Let us know →

Get weekly workflows — subscribe to the newsletter.