How to Connect HubSpot to Claude: Unlock AI CRM Insights in 1 Minute

Sales teams live in HubSpot. But when a VP asks, “Why did we miss quota last month?” or a marketer needs to know which email sequence is actually converting leads, the answer isn’t a click away; it’s buried in a CRM that was built to store data, not explain it.

HubSpot’s native reports are powerful but rigid. You can pull a deals report or a contact activity summary, but you can’t ask a follow-up question. You can’t request “actually, show me only the deals that stalled at the proposal stage” and get an instant answer. That kind of dynamic, conversational analysis requires either a dedicated analyst or hours of manual filtering.

That changes with Windsor MCP. Connect your HubSpot CRM to Claude in three simple steps and get AI-powered insights into your deals, contacts, companies, email sequences, and marketing activity in seconds.

🚀 Build your first HubSpot to Claude integration with a 30-day free trial: https://onboard.windsor.ai/.

Even better, you can instantly generate comprehensive visual reports from your data—like the example shown here.

Here’s how you get started.

3 steps to connect HubSpot to Claude via Windsor MCP

No developers needed. The setup takes less than a minute and runs on autopilot from there.

📖 Full step-by-step guide: https://windsor.ai/documentation/windsor-mcp/how-to-integrate-data-into-claude/.

Prerequisites

• A Windsor.ai account (free trial or paid plan)
• A HubSpot account
• A Claude.ai account

Step 1: Connect HubSpot in Windsor

1. Log in to onboard.windsor.ai.
2. Select HubSpot as your data source.
3. Authenticate with your HubSpot credentials.
4. Select the account(s) that you’d like to connect to Claude.

Once connected, Windsor pulls your CRM data directly from HubSpot’s API and structures it into a clean, LLM-optimized format. It normalizes deal stages, contact properties, company records, and email metrics under the hood, so Claude can read and reason across them without misinterpretation.

Step 2: Add the Windsor.ai connector in Claude

Open the direct Windsor.ai connector page and click Connect.

💡 Quick tip: Set permissions to “Always allow” to make your Windsor-integrated data available across every Claude conversation without re-authenticating.

Step 3: Start querying your HubSpot CRM data 

No SQL. No pivot tables. Just straightforward questions in human language, like these ones:

• “Which deals in the proposal stage have had no activity in the last 14 days?”
• “What’s our average deal velocity from first contact to closed-won this quarter vs. last quarter?”
• “Which sales rep has the highest close rate on enterprise deals?”
• “Show me contacts who opened our last email sequence but never booked a demo.”

💡 Pro tip: Ask Claude to combine pipeline health with marketing data and generate a visual report — something HubSpot’s native reports can’t do across objects without a custom report builder.

Bonus: HubSpot prompt ideas for Claude

These ready-to-use prompts are designed around the real questions sales and marketing teams usually ask:

Pipeline health check

Prompt:

Give me a full pipeline health summary as of today. How many open deals are in each stage? What is the total pipeline value? Identify any deals that have been in the same stage for more than 21 days and flag them as stalled.

Deal velocity analysis

Prompt:

Compare our average deal velocity (days from first contact to closed-won) this quarter vs. the previous two quarters. Is the sales cycle getting longer or shorter? Break this down by deal size — under $10k, $10k–$50k, and over $50k.

Rep performance benchmarking

Prompt:

Rank our sales reps by close rate for the last 90 days. For the bottom two performers, show me where in the pipeline their deals are most likely to stall — which stage has the highest drop-off rate for each rep?

Lead-to-customer conversion audit

Prompt:

Look at all contacts created in the last 60 days. What percentage have progressed to at least one sales activity? Of those, what is the conversion rate to demo booked, and then from demo to closed-won? Where is the biggest drop-off in our funnel?

Email sequence effectiveness

Prompt:

Analyze the performance of our active email sequences. Which sequence has the highest open rate? Which has the highest reply rate? Are there contacts who opened 3 or more emails in a sequence but never responded — and how many are there?

Churn risk identification

Prompt:

Look at our existing customer contacts. Which accounts have had zero activity logged — no calls, emails, or meetings — in the last 45 days? Cross-reference with deal size and flag any accounts over $20k ARR that may be at risk of going dark.

🤖 More prompt ideas in the Windsor AI Prompt Library: https://windsor.ai/prompt-library/.

What sales and marketing teams actually use HubSpot to Claude integration for

1. Revenue forecasting on demand

Instead of waiting for end-of-week pipeline reviews, you can ask Claude to generate a weighted revenue forecast at any moment, pulling live deal stages, close probabilities, and expected close dates from HubSpot.

2. Pre-meeting intelligence briefs

Before a client call, you can ask Claude to summarize a contact’s full HubSpot history: every interaction logged, their company details, any open deals, and the last email they engaged with. Get a concise brief in seconds rather than manually clicking through a contact record.

3. Marketing and sales alignment

Use Windsor MCP to combine HubSpot data with Google Ads or LinkedIn Ads in the same Claude chat. Ask: “Which paid campaigns are generating leads that actually close — and at what deal size?” This cross-source question is the one question that usually requires a BI tool and a data engineer to answer. With Windsor MCP, it’s a single prompt.

4. Onboarding and ramp analysis

For sales analysis, you can ask Claude to compare the pipeline progression patterns of your top performers against those of newer reps. Identify exactly where newer reps lose deals that experienced reps close, and build targeted coaching based on real data.

5. Automated CRM hygiene reports

Try asking Claude to identify contacts with missing required fields, deals with no close date set, or companies with no associated contacts. Use these insights to run a targeted CRM cleanup rather than waiting for a quarterly audit.

HubSpot data you can send to Claude via Windsor

Windsor streams a comprehensive set of HubSpot objects and properties (over 1000 metrics) into Claude, including the following:

CRM objects

• Contacts — Name, email, phone, lifecycle stage, lead source, creation date, last activity date, and all standard and custom contact properties
• Companies — Company name, domain, industry, revenue, employee count, associated contacts, and deals
• Deals — Deal name, stage, amount, close date, associated contact and company, deal owner, creation date, and custom deal properties
• Activities — Calls, emails, meetings, notes, and tasks logged against contacts and deals — the full interaction history

Sales pipeline metrics

• Deal stage — Current pipeline stage for every open and closed deal
• Close probability — HubSpot’s assigned probability by stage, used for weighted forecasting
• Deal velocity — Time spent in each pipeline stage and total days from creation to close
• Owner assignment — Which rep owns each deal, contact, or company

Marketing and email data

• Email sends, opens, clicks — Performance metrics per email and per sequence
• Email sequence enrollment — Which contacts are enrolled in which sequences and at what step
• Form submissions — Lead capture activity tied to specific landing pages or campaigns
• Marketing contact status — Whether a contact is a marketing contact and their subscription preferences

Lifecycle and attribution

• Lifecycle stage — Subscriber, lead, MQL, SQL, opportunity, customer, evangelist
• Lead source — Original source attribution: organic search, paid, social, email, referral, etc.
• UTM parameters — Campaign, medium, source, and content tags passed through from marketing campaigns

💡 Windsor normalizes HubSpot’s object relationships, so Claude understands that a deal is connected to a contact, which is connected to a company, which has email engagement history. This relational context is what makes cross-object questions possible.

Why most teams are still stuck connecting HubSpot to Claude manually

The irony of HubSpot is that it’s designed to centralize your customer data, yet getting answers out of it in a flexible way remains genuinely hard.

Here’s what the manual workflow typically looks like, and where it breaks down.

HubSpot’s native reports

HubSpot’s report builder is functional for standard views: pipeline by stage, deals closed by rep, and email open rates. But it’s object-constrained; a single report can’t easily pull a contact’s email engagement history alongside their deal stage and last activity date.

Cross-object analysis requires custom reports, which need Professional or Enterprise tier access and someone who knows the report builder well.

CSV exports into Claude

Some teams export individual object lists (a contacts CSV, a deals CSV ) and upload them separately to Claude.

The fundamental problem: HubSpot’s exports don’t preserve relationships between objects. You get a flat list of deals with no linked contact history, or a contact list with no pipeline context. Claude can analyze each file, but it can’t reason across them unless you manually combine them first.

HubSpot’s native AI features

HubSpot has been rolling out its own AI-powered features, such as ChatSpot, AI summaries, and forecasting tools. These are useful but operate strictly within HubSpot’s interface and data model. They can’t combine HubSpot data with your paid ad spend, your website analytics, or any other data source.

The moment you need a question answered that crosses platforms, you’re back to manual work.

Comparison: Top ways to get HubSpot data into Claude

Conclusion

HubSpot is where your most valuable business data lives: every prospect conversation, every email open, every deal won or lost. The problem has never been a lack of data. It’s that turning that data into a clear answer to a specific question has always required either deep HubSpot expertise, a custom report, or a spreadsheet someone built months ago and half the team doesn’t trust.

Connecting HubSpot to Claude via Windsor MCP doesn’t just make reporting faster. It changes who can ask the questions. A sales manager can interrogate their pipeline at 7 am before a board call. A marketer can trace which campaign touchpoints actually influenced closed revenue without filing a request with the data team. A founder can ask “what’s our average time to close for inbound vs. outbound leads this quarter” and get an answer before the coffee’s brewed.

That kind of access changes how revenue teams operate. And it starts with a 60-second setup when you use Windsor MCP.

🚀 Ready to start? Connect HubSpot to Claude with a 30-day free trial at Windsor.ai: https://onboard.windsor.ai/app/hubspot.

FAQs

Does Windsor support HubSpot Starter, Professional, and Enterprise?

Yes. Windsor connects to all HubSpot tiers. The data available to Windsor depends on what your HubSpot plan includes. For example, custom deal properties and advanced sequence data require Professional or Enterprise, but the Windsor connector itself works regardless of tier.

Can Claude query across multiple HubSpot objects at once?

Yes, and this is one of the key advantages over HubSpot’s native exports. Windsor preserves the relational structure between HubSpot objects (contacts linked to deals, deals linked to companies, activities linked to contacts). This means Claude can answer cross-object questions like “show me all contacts who are in an active email sequence and also have an open deal in the proposal stage” without you needing to manually join data.

What is the fastest way to get HubSpot data into Claude?

Windsor MCP is the fastest method by a significant margin. The setup takes under 60 seconds at onboard.windsor.ai and requires no code or ongoing maintenance. Once connected, your HubSpot data is available in every Claude conversation in real time — no exports, no file uploads, no waiting.

What are the main ways to connect HubSpot to Claude?

There are three available options:

• Windsor MCP: A no-code, live connector. Set up once, query forever. Preserves relational context between HubSpot objects. Works with 325+ other data sources in the same Claude chat.
• Manual CSV export: Download object lists from HubSpot and upload to Claude. Works for simple one-object questions, but loses relational context and requires repeating the process every time.
• Direct HubSpot API: Build a custom integration with full flexibility. Requires developer resources, HubSpot API knowledge, OAuth management, and ongoing maintenance as the API evolves.

For most sales and marketing teams, Windsor MCP is the only approach that’s fast enough for daily use and deep enough for real analysis.

Can I combine HubSpot data with my ad spend data in Claude?

Yes, this is one of the most powerful use cases for the Windsor + Claude combination. You can connect HubSpot alongside Google Ads, LinkedIn Ads, Facebook Ads, or any of Windsor’s 325+ connectors. Ask Claude questions that span both systems: “Which paid campaigns generated leads that actually closed — and what was the average deal size?” — the kind of question that normally requires a BI tool.

Does this work with HubSpot custom properties and custom pipelines?

Yes. Windsor pulls standard and custom properties from your HubSpot account, including custom contact properties, custom deal properties, and any custom pipeline stages you’ve configured. Claude can filter, group, and analyze by any property that exists in your HubSpot instance.

Can I use this for marketing analytics, or is it mainly for sales?

Both. Windsor streams HubSpot’s marketing data, including email sequences, form submissions, contact lifecycle stages, lead source attribution, and UTM parameters, alongside the sales pipeline data. Marketing teams use it for funnel analysis, email performance reviews, and campaign attribution. Sales teams use it for pipeline management, forecasting, and rep benchmarking. The most powerful use cases are the ones that connect both.

Is my HubSpot data secure when using Windsor?

Windsor is SOC 2 compliant and uses OAuth-based authentication; it never stores your HubSpot credentials. Data is transmitted securely, and Windsor requests only the read permissions necessary to pull your analytics data. See Windsor’s full security documentation at windsor.ai/security/.

Can Claude write or update records back into HubSpot?

No, Windsor MCP builds a read-only data pipeline. Claude can analyze, summarize, and reason about your HubSpot data, but it cannot create, update, or delete records in your CRM. This is intentional: it ensures your CRM data stays clean and changes are always made deliberately within HubSpot.

How is this different from HubSpot’s own AI features like ChatSpot?

HubSpot’s native AI tools are useful for in-app tasks: drafting emails, summarizing contact records, and basic forecasting, but they’re confined to HubSpot’s own interface and data model. They can’t combine your CRM data with your ad performance, website analytics, or any external source. Windsor MCP gives Claude access to HubSpot data alongside 325+ other sources, so you can answer cross-platform questions.