By Pindi Sahota · Last updated: 2026-06-07
This page contains affiliate links. If you purchase through them, I may earn a commission at no extra cost to you.
Claude for CRM Automation — Full Workflow Guide (2026)
Last updated: 2026-06-07
Claude CRM automation eliminates the manual work that clogs every sales pipeline: writing follow-up emails from scratch, scoring leads based on form data, researching prospects before calls, and updating records after meetings. Claude connects to your CRM through Make.com or n8n, reads contact and deal data, and outputs structured content or decisions that your automation routes back into the correct fields. The result is a CRM that stays up to date, with personalised outreach and qualified leads, without your sales team spending hours on admin. This guide covers which CRMs work with Claude, four automation examples, and how to build your first workflow.
How Claude Connects to Your CRM
Claude does not connect directly to CRMs — it has no native HubSpot or Salesforce plugin. The connection happens through an automation platform that sits between Claude and your CRM:
Via Make.com: Make.com has native modules for HubSpot, Salesforce, Pipedrive, Zoho CRM, ActiveCampaign, and Notion. You add the Anthropic (Claude) module between a CRM trigger and a CRM action. Make.com handles all API authentication and data mapping visually.
Via n8n: n8n has built-in nodes for HubSpot, Salesforce, Pipedrive, and Notion. Like Make.com, you chain the Claude HTTP Request node between CRM trigger and action nodes. n8n is better for high-volume scenarios or teams that need to self-host.
Via direct API code: If you're a developer, you can call both the CRM API and the Anthropic API directly in Python or Node.js. This gives maximum flexibility but requires maintenance.
For most teams, Make.com is the fastest path to a working Claude CRM automation.
CRM Compatibility with Claude
| CRM | Make.com module | n8n node | Native Claude | API access |
|---|---|---|---|---|
| HubSpot | Yes (native) | Yes (native) | No | Yes |
| Salesforce | Yes (native) | Yes (native) | No | Yes |
| Pipedrive | Yes (native) | Yes (native) | No | Yes |
| Notion | Yes (native) | Yes (native) | No | Yes |
| Zoho CRM | Yes (native) | Community node | No | Yes |
| ActiveCampaign | Yes (native) | Yes (native) | No | Yes |
| Close | Via HTTP | Via HTTP | No | Yes |
| Copper | Yes (native) | Via HTTP | No | Yes |
"Via HTTP" means no native node exists, but n8n's HTTP Request node can call the CRM's REST API directly. Any CRM with a REST API can connect to Claude via Make.com's HTTP module or n8n's HTTP Request node.
4 Claude CRM Automation Examples
Automation 1: Lead Scoring on Form Submission
Trigger: Make.com watches for new HubSpot contacts (or any form submission) Claude prompt: ` Score this new lead from 1-10 based on fit for our B2B SaaS product. High score (8-10): Company 50+ employees, B2B industry, decision-maker title, specific pain described. Low score (1-4): Individual/consumer, vague message, student email. Mid (5-7): Everything else.
Name: {{name}} Company: {{company}} Job title: {{title}} Company size: {{company_size}} Message: {{message}}
Respond with JSON only: {"score": number, "tier": "hot|warm|cold", "reason": "one sentence"} ` Action: HubSpot > Update Contact — writes score to a custom property, updates the lifecycle stage based on tier Action 2 (hot leads only): Slack notification to the sales channel: "New hot lead: {{name}} from {{company}} — Score: {{score}}"
Impact: Every lead enters your CRM pre-scored. Sales reps see who to contact first without reviewing each form submission.
Automation 2: Pre-Call Research Brief
Trigger: Make.com Schedule trigger (runs every morning at 7am) Step 1: HubSpot > List Contacts filtered by "meeting scheduled today" Step 2: For each contact, HubSpot > Get Contact (fetches full record: company, deal history, last activity, notes) Claude prompt: ` Prepare a 5-bullet pre-call brief for a sales rep meeting with this prospect today. Include: their likely goals, key concerns based on their form message, suggested opening question, relevant company context, and one personalisation point.
Contact record: {{contactdata}} Deal stage: {{dealstage}} Last activity: {{lastactivitynote}} ` Action: Gmail > Send Email to the assigned sales rep with the brief
Impact: Sales reps arrive at every call prepared, without spending 20 minutes researching each prospect. Response times to "how did the call go?" improve because the rep felt confident.
Automation 3: AI-Written Follow-Up Email Drafts
Trigger: HubSpot > Watch Deal Stage Changes (fires when a deal moves to "Follow Up Needed") Claude prompt: ` Write a 150-word follow-up email to {{contactname}} at {{company}}. Context: We just had an introductory call. They showed interest in {{productinterest}} but raised concern about {{objection}}. Goal: Keep the conversation warm, address the concern briefly, and propose a 30-minute demo next week. Subject line: Include a compelling subject line prefixed with "Subject:" Tone: Warm and confident, not pushy. ` Action: Gmail > Create Draft (saves as a draft in the rep's Gmail inbox — they review and send) Action 2: HubSpot > Create Task ("Review and send follow-up email draft") assigned to the deal owner
Impact: Reps get a first draft ready in their Gmail. They spend 2 minutes editing rather than 15 minutes writing from scratch. Follow-up rate increases because the friction of "I need to write this email" is removed.
Automation 4: Post-Meeting Note Summariser
Trigger: Webhooks > Custom Webhook (your meeting tool sends a transcript webhook — Otter.ai, Fireflies, or similar) Claude prompt: ` Summarise this sales call transcript into a CRM note. Include:
- Key topics discussed (3-5 bullets)
- Next steps agreed (bulleted list)
- Objections or concerns raised
- Proposed timeline / deal stage recommendation
Transcript: {{transcript}} ` Action: HubSpot > Create Note on the contact record (pastes Claude's summary) Action 2: HubSpot > Update Deal — update close date and stage based on Claude's recommendation (if Claude extracted that information)
Impact: CRM records update automatically after every call. No more half-filled deal records because reps didn't have time to write notes.
Setting Up Your First Claude HubSpot Automation in Make.com
Step 1: Connect HubSpot to Make.com
In Make.com, create a new scenario. Add a HubSpot trigger module. Click Create a connection and follow the OAuth flow to connect your HubSpot account. Choose your trigger type (Watch New Contacts, Watch Deal Stage Changes, etc.).
Step 2: Add the Anthropic Module
After the HubSpot trigger, add an Anthropic > Create a Message module. Create an API key connection using your Anthropic API key. Write your prompt in the Content field, mapping HubSpot data variables ({{name}}, {{company}}, {{message}}) into the prompt text.
Step 3: Parse Claude's Response
If Claude returns JSON (for scoring or structured data), add a JSON > Parse JSON module to extract the values into individual variables. If Claude returns plain text (for email drafts or summaries), you can map the raw text output directly.
Step 4: Update the CRM
Add a HubSpot action module — Update Contact, Create Note, Create Task, or Send Email. Map Claude's parsed output into the relevant HubSpot fields.
Step 5: Add Error Handling
Add an error handler to the Anthropic module (right-click > Add error handler). If the API call fails, set a fallback: send yourself a Slack message with the contact details so you can process it manually.