AI for HubSpot: Tools, Examples and Tips

AI for HubSpot starts with a familiar Monday. Twenty new leads from last week’s trade fair sit in a spreadsheet, half of the deals in the pipeline have not been touched for a month, the call notes from Friday are still in someone’s notebook, and a customer asks in the chat when the quote arrives. HubSpot holds all of it, but somebody still has to type it in, sort it and follow up.

This guide shows what HubSpot’s own AI does, where it stops, ten tools that work with HubSpot, and how teams hand over CRM work from Outlook, Teams or Slack.

What AI for HubSpot Is

AI for HubSpot is software that reads and updates your HubSpot CRM for you: it qualifies leads, enriches contacts, logs calls and emails, moves deals, answers customer questions and drafts the follow-up, so your team checks and approves instead of typing. It comes either from HubSpot itself or from tools that connect through HubSpot’s official API, its connectors for ChatGPT and Claude, or its MCP server.

How It Works with Superkind

Superkind is the AI employee for your team. You write to Superkind in Outlook, Gmail, Microsoft Teams, Slack or the Superkind web app, like you would message a colleague. Superkind is hosted in Europe, GDPR compliant and connects to 1,500+ integrations, so it does the task in HubSpot and the tools around it, such as Outlook, Excel, DocuSign and DATEV, and asks you before anything goes out. See the HubSpot integration for what it can do there.

1. 1

   You write the task

   Write in Teams or forward an email from Outlook: “Qualify yesterday’s trade fair leads in HubSpot.” No new screen, no workflow builder. You ask in plain language.

2. 2

   Superkind does it in HubSpot

   It creates or updates the contacts and companies, checks each lead against your criteria, sets the lifecycle stage, assigns the owner and drafts the first email, using your rules and the Company Brain: your customers, prices and processes.

3. 3

   It asks for approval

   Before an email goes to a prospect or a deal changes stage, Superkind shows you the result with Approve or Reject, in Teams, Slack or Outlook, wherever you are. A rule like “Quotes over EUR 10,000 need sign-off from management” sends the approval to the right person.

4. 4

   It becomes a routine

   What you explain once keeps running: “Every Monday, list the deals with no activity for 30 days and draft a follow-up for each.” The result lands where your team already works.

From the First Task to Your Own App

Superkind grows with your HubSpot work in four stages. Each one builds on the last. How this looks for a whole sales team is on the AI sales rep page.

1. 1

   Ready in minutes

   Sign up for free, connect HubSpot, send your first lead list in Teams, Slack or by email and watch Superkind enter and qualify it.

2. 2

   Tasks become routines

   Lead qualification, call notes and the Monday pipeline check run as routines, and Superkind asks for approval before any email goes to a customer.

3. 3

   Learns from day one

   Every task teaches Superkind your ideal customer, your price list and who owns which region, and this Company Brain is never lost, even when a rep leaves.

4. 4

   When it gets more complex, you do not start over

   Build your own quote-to-order app that connects HubSpot deals with DocuSign and DATEV, on top of what Superkind already knows: your own Super App, built yourself or together with us. It all belongs to you.

What HubSpot Brings Itself

HubSpot has built a lot of AI into its platform, and for many teams it is the first place to look. In July 2026 HubSpot regrouped its agents: Breeze Agents and Breeze Studio became Agent Hub and Agent Builder.³ This is what you get today:

• Breeze Assistant - the AI assistant inside HubSpot for writing, summarising records and questions about CRM data. It kept its name in 2026.⁴
• Agent Hub - one place for every ready-made and custom agent in your portal, with usage across agents. It is listed for Professional and Enterprise and runs in public beta.³
• Agent Builder - build your own agents on CRM data. You can describe what you want in plain language and HubSpot sets up instructions, actions and knowledge sources for you to edit.³
• Customer Agent - answers customer questions in chat and other channels from your knowledge base and hands over to a person when needed. Since 14 April 2026 it costs 50 credits (USD 0.50) per resolved conversation.¹
• Prospecting Agent - researches contacts and recommends leads for outreach. It costs 100 credits (USD 1) per lead recommended for outreach.¹
• Credits - agents and many AI features use HubSpot Credits. Extra capacity costs USD 10 per 1,000 credits.² Custom agents cost credits per run, depending on the tokens used.³
• MCP server and connectors - HubSpot offers an official MCP server that gives MCP-capable AI tools read and write access to CRM data with OAuth,⁵ plus its own connectors for ChatGPT and Claude.^6,7
• UNBOUND 2026 betas - in September 2026 HubSpot showed new agents in public beta, such as a Campaign Agent and a Content Agent.⁴

For teams on Professional or Enterprise who live in HubSpot all day, this covers a good share of CRM work. Prices as of October 2026.

Where the Gap Is

HubSpot’s AI is built for work in HubSpot. Much of the work starts and ends somewhere else:

• Agents work inside HubSpot - the lead arrives in Outlook, the price list sits in SharePoint, the contract in DocuSign, the invoice in DATEV. A HubSpot agent sees the CRM, not the work around it.
• Professional or Enterprise only - Agent Hub and Agent Builder are listed for Professional and Enterprise.³ Teams on Free or Starter do not get Agent Hub.
• Costs per run are hard to plan - outcome pricing helps for the Customer and Prospecting Agents, but custom agents cost credits per run depending on tokens, so you need limits per agent.³
• No approval where people are - the review happens in HubSpot. A sales lead who lives in Outlook and Teams does not get an Approve button there before the email goes out.
• Chat connectors answer, they do not finish the job - ChatGPT and Claude answer pipeline questions well, but each person needs a subscription and copies the result onwards.
• Knowledge beyond the CRM - your rules (“Complaints go to customer service”, “Quotes over EUR 10,000 need sign-off”) and what your best rep knows rarely sit in HubSpot properties.

An AI employee fills this gap: HubSpot stays the system of record, and the work around it gets done too.

10 AI Tools for HubSpot at a Glance

All ten tools work with HubSpot, each for a different job.

The 10 Tools in Detail

1. Superkind

Superkind is one AI employee that works where your team works: Outlook and Gmail, Microsoft Teams, Slack and the web app. For HubSpot it qualifies leads, logs call notes, keeps deals current, sends quotes, triages tickets and builds the pipeline report.

• Works in HubSpot and beyond - contacts, companies, deals, tickets and activities in HubSpot, plus Outlook, Excel, SharePoint, DocuSign and DATEV.
• Asks before anything goes out - Approve or Reject before an email is sent or a deal changes, in the channel where the person is.
• Company Brain and rules - it knows your ideal customer, prices and owners and applies your rules, so the knowledge stays when people leave.
• Routines and custom apps - explain a task once and it runs on schedule; when it grows, build your own apps on the platform with 1,500+ integrations.

Good fit for:

• Sales, marketing and service teams that work in Outlook and Teams more than in HubSpot
• Teams on any HubSpot edition, including Starter
• Work that touches HubSpot and other systems
• Anyone who wants an approval step before customers get an email

2. HubSpot Breeze Assistant and Agent Hub

HubSpot’s own AI is the natural choice when the work happens in HubSpot. Breeze Assistant helps every user write and summarise, and Agent Hub holds ready-made agents such as the Customer Agent and the Prospecting Agent, plus your own agents from Agent Builder.^1,3

• Native CRM context - agents see the records, properties and history in your portal.
• Outcome pricing - the Customer Agent bills per resolved conversation, the Prospecting Agent per recommended lead.¹
• Agent Builder - describe an agent in plain language and edit what HubSpot sets up.³
• Credit controls - monthly credit limits per agent and usage in one view.³

Good fit for:

• Teams on Professional or Enterprise
• Service teams with a solid knowledge base
• Marketing teams that build campaigns in HubSpot
• Work that starts and ends in HubSpot

3. ChatGPT with the HubSpot connector

HubSpot’s connector for ChatGPT lets users ask questions about contacts, companies, deals and tickets in ChatGPT, in chat or deep research mode. A HubSpot Super Admin has to approve it first, and each user only sees the data their HubSpot permissions allow.⁶

• Chat on CRM data - quick questions like “Which deals close this month?”
• Deep research - longer analysis across CRM data and the web.⁶
• Admin approval - a Super Admin decides which data the connector may use.⁶
• Permissions respected - users see only what they may see in HubSpot.

Good fit for:

• People who already work in ChatGPT
• Pipeline analysis and account research
• Preparing for customer meetings
• Ad-hoc questions without building reports

4. Claude with the HubSpot connector

HubSpot built a connector for Claude that answers questions about CRM data in natural language and can create charts. It is available to HubSpot customers on all tiers with a paid Claude plan; HubSpot states that shared data is not used for training unless users give feedback or opt in.⁷

• Natural-language queries - ask about deals, contacts and tickets.
• Charts and analysis - turn pipeline data into a quick visual.
• Permission-aware - a rep sees only the deals they may see.⁷
• OAuth setup - install from Claude’s connector settings and sign in to HubSpot.

Good fit for:

• Managers who want answers without building reports
• RevOps analysis
• Teams already on Claude Team or Enterprise
• Meeting preparation

5. Zapier

Zapier connects HubSpot with thousands of apps through triggers and actions, and you can add AI steps to a Zap, for example to sort an inbound lead.

• HubSpot triggers - new contact, deal stage change, form submission.
• AI steps - classify, summarise or extract data inside a flow.
• Large app catalogue - connect HubSpot to tools the native integrations miss.
• No code - flows are built in a visual editor.

Good fit for:

• Teams with a clear, fixed process to automate
• Someone who likes to build and maintain flows
• Simple handovers between HubSpot and other apps
• Notifications and data sync

6. Make

Make builds automation as visual scenarios with HubSpot and AI modules. You can choose an EU data zone; Make notes that some AI features are still processed in the US or the UK.⁹

• Visual scenarios - branches, filters and loops on one canvas.
• HubSpot modules - create, update and search CRM records.
• AI modules - add language models to a scenario.
• Data zones - US or EU zones at organisation level.⁹

Good fit for:

• Ops teams with complex, rule-based flows
• People comfortable with logic and data mapping
• Teams that want an EU data zone for automation
• Sync between HubSpot and back-office tools

7. Clay

Clay enriches contacts and companies from many data providers, runs AI research on each record and writes the results back to HubSpot.

• Waterfall enrichment - checks several data sources until it finds a value.
• AI research - reads websites and fills custom fields per account.
• HubSpot write-back - creates and updates records in the CRM.
• Table workflows - a spreadsheet-like view for lists.

Good fit for:

• Outbound teams that build account lists
• RevOps teams cleaning up CRM data
• Segmentation by firmographics
• Personalised first lines for outreach

8. Gong

Gong records and analyses sales calls and emails, flags deal risks and syncs activities to HubSpot.

• Call recording - transcripts and highlights of customer calls.
• Deal insights - signals on deal risk and next steps.
• Coaching - managers review calls and give feedback.
• HubSpot sync - activities and insights on the CRM record.

Good fit for:

• Sales teams with many customer calls
• Sales leaders who coach reps
• Forecast reviews
• Longer B2B sales cycles

9. Aircall

Aircall is cloud telephony that logs calls in HubSpot. With its AI add-on it can write call summaries into the HubSpot call notes, visible on the contact, company or deal.

• Click-to-call - call from HubSpot records.
• Automatic logging - calls appear on the right record.
• AI summaries - call summaries in HubSpot call notes.
• Team routing - queues and routing for sales and support.

Good fit for:

• Inside sales teams
• Support teams on the phone
• Teams that want call notes without typing
• Companies switching to cloud telephony

10. Fathom

Fathom records Zoom, Google Meet and Teams meetings and writes summaries and action items. Its HubSpot app syncs them to meetings, tasks and deal fields.

• Meeting summaries - notes ready right after the call.
• Action items - follow-ups as HubSpot tasks.
• Deal field mapping - summaries mapped to deal properties.
• Works with video tools - Zoom, Google Meet and Microsoft Teams.

Good fit for:

• Reps with many video calls
• Founders who sell themselves
• Customer success check-ins
• Teams that forget to log meetings

Typical Use Cases

These HubSpot tasks eat the most time, and AI helps fastest here.

Qualifying inbound and trade fair leads

The AI checks each new lead against your criteria, enriches the company, sets the lifecycle stage and assigns the owner. With Superkind you send the list in Teams and approve the follow-up emails before they go out. The step-by-step version is in our guide to qualifying leads with AI in HubSpot.

Logging calls and meetings

Call notes that never reach the CRM make pipelines look better than they are. Aircall, Fathom or Gong write summaries into HubSpot; Superkind logs a short message from Teams on the right deal.

Keeping deals and contacts current

Stale deals, duplicate contacts and missing fields distort every forecast. A weekly routine finds deals without activity, flags duplicates and proposes updates for the owner to approve.

Sending quotes

Superkind drafts the quote from your Excel price list, attaches it to the HubSpot deal, asks for approval above your threshold and sends it from Outlook.

Triaging service tickets

HubSpot’s Customer Agent answers common questions in chat. For the rest, an AI employee sorts tickets by urgency, routes complaints to customer service and drafts the reply for approval.

Building the weekly pipeline report

Set it as a routine: “Weekly report to Daniel every Friday”, with new, won and lost deals and the top risks, delivered in Teams or by email.

Benefits of AI for HubSpot

The goal is a CRM your team trusts because it is complete, without anyone typing all afternoon.

A CRM that stays complete

When calls, emails and meetings land in HubSpot automatically, forecasts and reports reflect what actually happens. Our article on CRM data decay shows what stale data costs.

Faster follow-up

AI qualifies the lead and drafts the reply the same day; a person only approves it.

Less switching between tools

Reps live in Outlook and Teams. When they hand over HubSpot work from there, the CRM still gets updated.

Knowledge that stays

Which customer gets which discount and who owns which region usually lives in a few heads. Superkind keeps it in the Company Brain and applies it through rules, so a new rep inherits a system that already knows the business.

Control before anything goes out

Good AI for HubSpot proposes, you approve. Customer emails, stage changes and quotes above a threshold wait for a person, and every action is logged.

Data Protection and EU Hosting

HubSpot holds names, email addresses, phone numbers and the full history of every customer conversation. Any AI that reads or writes this data processes personal data under the GDPR. These are the facts for Superkind:

Safe by default

Your data stays in Europe, under a DPA, with no training on your data.

• Hosted and processed in Europe
• GDPR compliant
• Aligned with the EU AI Act
• No training on your data
• DPA through the Trust Center

Superkind uses AI models from OpenAI, Anthropic, Mistral and Google, processed in Europe. HubSpot itself offers an EU data centre in Frankfurt for customer data;⁸ check its subprocessor list for where AI features run. For the other tools, the table shows what each vendor states; where it says “check with vendor”, ask for the hosting location and the DPA before you connect your CRM.

6 Tips for Using AI with HubSpot

1. Start with one task that annoys everyone

Pick the task your team complains about most, usually lead qualification or logging calls, and get it running end to end. One visible win earns more trust than five pilots.

2. Clean up the properties first

AI fills the fields you have. If lifecycle stages, deal stages and owners mean different things to different people, write down the definitions before you hand over the work. HubSpot’s new Context Home scores how ready your data is for AI.⁴

3. Write your rules down

“A leads go to Daniel”, “Complaints go to customer service”, “Quotes over EUR 10,000 need sign-off from management”. Rules like these turn an AI that guesses into one that works like your best rep.

4. Keep approval on customer-facing steps

Let AI prepare the email, the quote and the stage change, but keep a person on the Approve button before anything reaches a customer. Loosen it later for routine cases.

5. Sort out data protection before connecting

Sign the data processing agreement (for Superkind it is in the Trust Center), check where the data is processed and give the AI only the HubSpot permissions it needs. Under Article 4 of the EU AI Act¹⁰, companies using AI must make sure their staff have enough AI literacy, so brief the team on what the AI does and where they approve.

6. Set a budget and watch usage

Credits are spent per conversation, lead or run. Set monthly limits per agent in HubSpot,³ and compare with a flat company price. Superkind starts free and Business starts from EUR 50 per month for unlimited users (as of October 2026).

Conclusion

HubSpot’s own AI has grown up fast: Breeze Assistant for everyone, Agent Hub and Agent Builder on Professional and Enterprise, outcome pricing and an official MCP server. It works best for work that starts and ends in HubSpot. The work around the CRM, in Outlook, Teams, Excel, DocuSign and DATEV, needs an AI that works across systems and asks before anything goes out.

The quickest way to see the difference is one real task. Send Superkind your next lead list in Teams or Outlook, let it qualify the leads in HubSpot through its 1,500+ integrations, approve the follow-up emails, and turn it into a routine. Start for free, or book a demo to see it run in your own HubSpot.

Frequently Asked Questions

Related Articles

• How to Qualify Leads with AI in HubSpot: Step-by-Step Guide
• The Best AI-Powered CRM Tools in 2026: An Honest Buyer Comparison
• When the CRM Lies: The Hidden Cost of Stale Sales Data
• AI in Sales: How the Mittelstand Automates Lead Qualification, Proposals, and CRM Hygiene

Henri Jung

Co-founder of Superkind, the AI employee teams write to in Outlook, Teams and Slack. Henri helps companies, from small teams to the Mittelstand, hand their routine work to AI - without new tools, consultants or an in-house AI team.