From theory to action: how real teams use AI inside their CRM

The HubSpot + ChatGPT deep research connector brings artificial intelligence directly into your CRM, giving every team the ability to ask real questions about their data and get clear, actionable answers.

No more exporting spreadsheets or waiting on reports. Just open ChatGPT, enable HubSpot as a source, and type your question in plain English.

To help you get started, here are real-world prompt examples for marketing, sales, and service teams, with sample outputs you can copy, paste, and build on inside your own HubSpot portal.

How It Works

Once you’ve connected HubSpot to ChatGPT, you can use two main modes:

• Chat Mode for quick lookups (“Show me all deals closing this month”).
• Deep Research Mode for multi-step analysis (“Analyze which deal sources had the highest close rates over the last three quarters”).

Both modes use read-only CRM access, meaning your data stays secure while ChatGPT analyzes contacts, companies, deals, and tickets based on your permissions.

Each of the following prompt examples is designed to help you:

• Ask better questions of your CRM data.
• See insights that were previously buried in reports.
• Turn those insights into immediate action inside HubSpot.

10 Prompts To Get You Started Using HubSpot + ChatGPT

Prompts for Marketing Teams

Use ChatGPT to identify your best-performing audiences, campaigns, and channels, without manual data crunching.

Prompt 1:

“In HubSpot, find my highest-converting contact segments from the past 90 days. Group by source and industry, and summarize which campaigns generated the most qualified leads.”

Sample Output:

Your top-converting contacts come from the “Webinar Registration” source, primarily in the manufacturing and SaaS industries. The campaigns driving the most qualified leads were “Spring Webinar Series” (4.8% conversion) and “Ebook: Modernizing CRM” (3.9%). Consider replicating webinar follow-up workflows for future campaigns.

Prompt 2:

“Analyze marketing email performance from the last three months. Which subject lines had the highest open and click rates by audience segment?”

Sample Output:

Emails with personalized subject lines (“[First Name], Ready to Automate Your CRM?”) outperformed others by 18%. B2B segments responded better to ‘how-to’ content, while B2C segments preferred list-style formats. Suggest A/B testing two new variations of your top performer next month.

Prompt 3:

“Identify which lead sources in HubSpot are producing contacts that become customers within 60 days.”

Sample Output:

Organic search leads convert fastest (median 44 days) and close 28% higher than paid ads. Partner referrals show the highest deal value but the slowest velocity. Recommendation: adjust lead scoring to favor organic and partner-sourced leads.

Try This Next:
Use your ChatGPT insights to build smart lists or automated workflows in HubSpot that nurture top-performing segments.

Prompts for Sales Teams

Use ChatGPT to find warm opportunities, diagnose pipeline risk, and uncover deal insights you might otherwise miss.

Prompt 1:

“Segment my target companies in HubSpot by annual revenue, industry, and tech stack. Identify top expansion opportunities based on deal history.”

Sample Output:

Companies using Salesforce and HubSpot show the strongest expansion potential (average deal size 2.3x higher than baseline). Recommended accounts for upsell: Acme Systems, GlobeTech, and DataBridge.

Prompt 2:

“Summarize the top reasons deals were lost last quarter. Include common competitor mentions and objections.”

Sample Output:

The three most common reasons for loss: 1) Budget freezes in Q2 (28%), 2) competitor pricing advantage (24%), and 3) missing integration with Slack (19%). Consider updating battlecards and positioning decks to highlight total cost of ownership and Slack integration roadmap.

Prompt 3:

“List open deals closing this month with probability above 70%, grouped by owner.”

Sample Output:

John Smith: 6 deals ($420K total), Maria Lopez: 4 deals ($310K total). Average close probability: 76%. Pipeline confidence is strong, forecast total close value $730K for August.

Try This Next:
Turn your insights into action:

• Create a HubSpot view or list for “Expansion Opportunities.”
• Use Breeze Customer Agent to automatically notify owners of high-priority accounts.

[eBook] AI in HubSpot: Claude, ChatGPT and Breeze Explained

Prompts for Service Teams

Use ChatGPT’s deep research capabilities to uncover patterns in support data and improve your customer experience.

Prompt 1:

“Analyze HubSpot tickets from the last 90 days. Identify the most common categories and the average resolution time for each.”

Sample Output:

Top categories: Billing (31%), Login Access (22%), and Onboarding (19%). Average resolution time is 7.4 hours, with 15% of tickets breaching SLA in the Billing category. Recommendation: expand automated responses for billing FAQs and reassign ticket ownership based on category expertise.

Prompt 2:

“Find all companies with recurring ticket issues related to the same product or feature.”

Sample Output:

Three accounts, BlueWave, NexusCorp, and InnovateX, show repeated issues with the API integration module. Suggest scheduling proactive outreach and updating documentation for API setup.

Prompt 3:

“Summarize support volume by week and forecast expected ticket volume next quarter based on historical seasonality.”

Sample Output:

Average weekly ticket volume has grown 11% since last quarter, with peaks during product releases. Forecast: 1,800 total tickets next quarter. Recommendation: increase part-time coverage by 20% during release cycles.

Try This Next:
Feed your findings into HubSpot’s Service Analytics or use Breeze Customer Agent to automate first-response messages during high-volume periods.

Best Practices for Getting Better Results

1. Always include “HubSpot” in your prompt. This ensures ChatGPT uses the connected CRM source.
2. Add context. Include filters like timeframe, team, or property names for more specific insights.
3. Keep your data clean. The quality of ChatGPT’s analysis depends on your CRM hygiene, duplicate records or outdated properties can affect accuracy.
4. Validate outputs. Review citations and source links back in HubSpot before acting on insights.
5. Use Deep Research for trends; Chat for quick lookups. Save time by matching your prompt to the type of answer you need.

The Takeaway

The HubSpot + ChatGPT connector turns everyday CRM data into actionable intelligence for every team. Whether you’re a marketer identifying your best-performing campaigns, a salesperson uncovering expansion opportunities, or a service leader forecasting ticket volume, you now have instant access to insights that used to take hours to uncover.

And because it’s built with read-only, permission-based access, you can explore your CRM confidently knowing your data stays secure.

Related Resources

AI Meets CRM: How HubSpot + ChatGPT Unlock Deep Research and Smarter Insights
HubSpot + ChatGPT: What It Can (and Can’t) Do With Your CRM Data
Deep Research vs Chat Mode: When to Use Each in HubSpot + ChatGPT
How to Connect HubSpot to ChatGPT (and What to Expect)
Data Privacy Deep Dive: What HubSpot + ChatGPT Can (and Can’t) See
Smarter Segmentation & Campaign Planning with HubSpot + ChatGPT
AI-Powered Prospecting: How Sales Teams Use ChatGPT Inside HubSpot
Retain and Renew: How Success Teams Use ChatGPT to Spot At-Risk Accounts
Predict Demand and Improve Support with ChatGPT Deep Research
AI for Data-Driven Operations: Using ChatGPT to Audit and Clean Your HubSpot Data
From Data to Decisions: How Executives Use ChatGPT for HubSpot Insights