This HubSpot AI guide explains what HubSpot’s AI (branded Breeze) does in 2026 and how to set it up step by step. Breeze is AI built into HubSpot’s Smart CRM. It answers questions about your records, drafts emails and reports, and runs agents that reply to customers, research leads and clean data.
Because Breeze reads your own CRM data, its output fits your business better than a separate chatbot. You will learn which agents each team should use, how HubSpot Credits are charged, what a real monthly budget looks like and which mistakes to avoid.
Quick Answer
HubSpot AI, branded Breeze, combines Breeze Assistant, task agents in Agent Hub and CRM-connected tools to automate marketing, sales and service work. Breeze Assistant comes with every HubSpot subscription at no extra cost. Most HubSpot-built agents need a Professional or Enterprise plan and are paid for with HubSpot Credits.
Key Takeaways
- Breeze Assistant is included with every HubSpot subscription at no extra cost.
- Agents now live in Agent Hub, and custom agents are built in Agent Builder (both public beta since July 23, 2026).
- Agents spend HubSpot Credits: 50 per resolved Customer Agent conversation, 100 per Prospecting Agent lead, $0.01 per extra credit.
The HubSpot Breeze Lineup at a Glance
Here is every part of HubSpot AI guide in one table (as of October 2026):
| Component | Primary Purpose | Best For |
|---|---|---|
| Breeze Assistant | Contextual assistance, CRM answers, drafts and reports | Individual users |
| Agent Hub (formerly Breeze Agents) | Running and managing agents that complete specialized tasks | Revenue and service teams |
| Agent Builder (formerly Breeze Studio) | Building custom agents | Admins and ops teams |
| Embedded AI Features | Writing, reporting and automation | Daily workflows |
| Smart CRM | Business and customer context | Connected AI output |
| AI Search Grader | Free one-time check of AI-search visibility | Marketing and AEO teams |
Verdict: HubSpot Breeze is best for teams that want AI to work directly with their CRM data, rather than through a separate general-purpose chatbot that lacks access to their pipeline, tickets or contact history.
What Is HubSpot AI?
HubSpot AI is the collection of artificial-intelligence capabilities now branded as Breeze. It works across marketing, sales, customer service, content and data operations, drawing on HubSpot CRM records, knowledge-base content, conversations and other business context to produce more relevant output than a standalone tool could.
Breeze is built to help your team, not to replace human review. Breeze Assistant is included with every HubSpot subscription at no additional cost, according to HubSpot. Specialized agents usually need an eligible subscription tier and HubSpot Credits to run at scale.
What Is HubSpot Breeze AI?
HubSpot Breeze is the AI layer inside HubSpot’s CRM. It answers questions about your records, drafts emails and reports, and runs agents that handle tasks such as support replies, prospect research and data cleanup. Everything it does pulls from your own contacts, deals, tickets and content, so the output fits your business. Breeze is really four layers working together:
- Breeze Assistant
- The conversational, user-directed layer for quick answers, summaries, drafts, proposals and reports.
- Agents in Agent Hub
- Task-oriented AI that completes defined jobs like prospecting or customer support with less manual input.
- Embedded AI features
- Smaller AI touches built into everyday tools: reporting, forecasting, email subject lines and more.
- Smart CRM context
- The underlying data layer that gives every one of the above its business-specific grounding.
For the most current specifications, HubSpot’s own ai product overview is the authoritative source and worth bookmarking, since AI features here update often.
Why Teams Use HubSpot AI Instead of a Separate Chatbot
Teams usually pick Breeze over a separate AI tool for three reasons:
- Tool consolidation
- One CRM-connected layer can cover writing, research and reporting, so you pay for fewer separate tools.
- Faster time-to-context
- Because Breeze already has access to deal stages, ticket history and past campaigns, it skips the manual copy-paste step that slows down general chatbots.
- Predictable governance
- Permissions, credit tracking and human-review settings live inside the same admin panel teams already use for the rest of HubSpot, which simplifies oversight compared to managing a separate AI vendor.
Breeze still needs a person to check its work. What it saves is the time spent gathering context and writing first drafts.
How HubSpot Breeze Works: Assistant, Agents and Smart CRM
The workflow behind every Breeze interaction follows a consistent pattern:
- HubSpot gathers permitted CRM, content and conversation context.
- A user submits a request, or an agent is triggered.
- Breeze analyzes the relevant business data.
- It drafts content, recommends an action or completes an approved task.
- Results are logged in HubSpot for review and reporting.
Breeze Assistant vs. Agents: Which Should You Use?
| Feature | Breeze Assistant | Agents in Agent Hub |
|---|---|---|
| Interaction | User-directed | Task-oriented or automated |
| Main role | Answers and assistance | Completes specialized work |
| Examples | Summaries, research, drafting, meeting prep | Prospecting, customer support, data cleanup |
| CRM context | Yes | Yes |
| Credit use | Included at no extra cost | Most agents use HubSpot Credits per outcome or action |
| Human review | Recommended | Configure approval and handoff rules |
A simple way to choose: if a person starts the task and reads the result, use Breeze Assistant. If the task repeats often, follows clear rules and you can measure the outcome (a resolved ticket, a researched lead, a cleaned record), an agent is worth testing. If you are unsure, start with the assistant. It costs nothing extra and shows you where the repeat work really is.
Breeze Intelligence and Data Enrichment
Breeze Intelligence was the name HubSpot used for its data enrichment layer. HubSpot’s current help docs simply call it data enrichment. It fills in missing company details such as industry, size and revenue, and shows website activity on contact and company records. Per HubSpot’s knowledge base (updated July 10, 2026), enrichment itself does not use HubSpot Credits. It is available on Starter, Professional and Enterprise of Marketing, Sales, Service, Data and Content Hub. Smart properties and buyer-intent tracking do use credits. Before you buy credits for enrichment, check what your current tier already includes.
How to Use HubSpot AI for Marketing, Sales and Service
This section shows which agents and tools each team actually uses.
HubSpot AI for Marketing
HubSpot AI marketing tools center on content and reporting speed:
- Content Agent (beta) drafts blog posts, social posts and landing pages from a brief.
- Campaign Agent (beta) plans campaigns and can be started from Breeze Assistant.
- Nurture Agent (beta) personalizes nurture emails for each contact.
- Social-media content assistance
- Landing-page and website generation
- Campaign reporting and predictive insights
- SEO and AI-search visibility tools
Teams that mainly need email campaigns rather than a full CRM can compare these features with mailchimp’s ai email writing and marketing automation tool. Teams that need the sending side without the CRM around it usually compare standalone email marketing platforms instead, and ecommerce brands in particular tend to narrow that search down to a head-to-head between omnisend and klaviyo. Pairing your CRM with other free tools? Here’s a full free marketing tool stack.
HubSpot AI for Sales
On the sales side, HubSpot’s agents and automation show up as:
- Prospecting Agent researches leads, tracks buying signals and drafts outreach.
- Deal Loss Agent reviews closed-lost deals and surfaces loss patterns and recommendations.
- Company Research Agent (beta) builds a summary of a target company before a call.
- Deal Progression turns call transcripts into CRM updates, follow-ups and deal plans.
- Personalized outreach drafts
- Conversation intelligence
- CRM summaries
- Suggested next actions
- Buying-signal monitoring
Reps should still verify AI-drafted messages and recommendations before sending or acting. Sales teams comparing specialist CRM platforms can also explore pipedrive’s ai sales assistant and automated follow-up features, including meeting summaries and deal management. Breeze speeds up the first draft, not the final judgment call. Teams evaluating general-purpose assistants alongside HubSpot’s sales tools often compare options like ChatGPT or Claude for open-ended writing and research that falls outside CRM-specific tasks.
HubSpot AI for Customer Service
Service teams get:
- Customer Agent answers text support conversations using your knowledge base. Voice support is in beta.
- Knowledge Base Agent (beta) writes new help articles where answers are missing.
- Customer Health Agent (beta) assesses account health from CRM activity, call transcripts and public information.
- Automated responses
- Ticket summaries
- Knowledge-gap identification
- Replies in more than one language (check language support for your agent before launch)
- Human-agent handoffs
If you are weighing a support-only AI agent instead, our Zendesk vs Intercom comparison covers the two biggest helpdesk options.
HubSpot AI for Data and Operations
Behind the scenes, operations teams use:
- Data Agent answers a prompt for each CRM record you point it at.
- Data Studio (beta) for building datasets used in workflows, segments and syncs
- CRM field population
- Customer research
- Workflow assistance
- Data enrichment
- AI-generated reports
- CRM record summaries
- Connected-app context
Feature matrix:
| Feature | Marketing | Sales | Service | Data/Operations |
|---|---|---|---|---|
| Content generation | Yes | Yes | Yes | No |
| Prospect research | No | Yes | No | Yes |
| Customer Agent | No | No | Yes | No |
| Predictive insights | Yes | Yes | Yes | Yes |
| Workflow automation | Yes | Yes | Yes | Yes |
HubSpot AI Search Grader (Free AEO Check)
AI Search Grader is HubSpot’s free, one-time check of how a brand appears in AI-generated answers rather than in traditional search rankings. According to HubSpot’s grader page (October 2026), it checks ChatGPT, Perplexity and Gemini, and gives a score out of 100 built from:
- Sentiment (up to 40 points), so yes, it does analyze brand sentiment across AI answers
- Presence quality (up to 20 points)
- Brand recognition (up to 20 points)
- Share of voice against competitors (up to 10 points)
- Market competition (up to 10 points)
For ongoing tracking and prioritized recommendations, HubSpot has a separate paid product called HubSpot AEO (beta). Run the free grader first to get a baseline before paying for tracking. Our list of ai seo tools for small businesses covers other ways to track AI-search visibility.
How to Set Up HubSpot AI in 6 Steps
Step 1: Identify your current hub and subscription tier, then confirm which Breeze features and agents are included at that level. Several of HubSpot’s own agents are limited to higher tiers.
Step 2: A Super Admin turns on generative AI access in AI Settings. Review account-level AI access, join a public beta only if a feature requires it, and assign appropriate user permissions.
Step 3: Clean up CRM records, connect approved knowledge-base articles, and review website, document and connected-app access. Breeze output is only as good as the data behind it.
Step 4: Good first matches: Customer Agent for repeat support questions, Prospecting Agent for lead research, Data Agent for filling gaps in records and Content Remix for turning one blog post into social posts and emails. Pick the task where you can count results within a month
Step 5: Provide audience details, define tone and terminology, and add escalation and approval rules before anything runs unattended.
Step 6: Start with a limited dataset, review accuracy, measure time saved and completed outcomes, and monitor credit consumption closely during the trial period. If you buy at least one Professional or Enterprise Service seat, HubSpot includes 14 days of free Customer Agent access.
Before You Switch On an Agent: Checklist
- Confirm your tier and seat type include the agent.
- Check whether the agent is in beta. Beta agents can change, and HubSpot says beta features may start using credits, with a goal of 30 days’ notice.
- Look up the agent’s credit rate and estimate one month of use.
- Set a feature-level credit limit for that agent.
- Choose who receives usage alerts.
- Pick one success number, such as resolution rate or meetings booked.
- Name one person who reviews agent output every week.
- Write down how to turn the agent off if results drop.
How HubSpot Credits Work (and How to Stay in Budget)
HubSpot Credits are the shared balance that pays for AI agents and other usage-based features. Each paid HubSpot account gets a set number of credits every month based on its highest tier. Agents use credits per outcome or action, unused credits expire at the monthly reset, and extra credits cost $0.01 each (as of October 2026).
Breeze has no single standalone price. What you pay depends on your hub, tier, seat count, contact volume and credit use. Here is what matters when you budget:
- Unused monthly credits do not roll over.
- Breeze Assistant is included at no extra cost, and data enrichment does not use credits.
- Agents, AI actions in workflows, buyer intent and some other features draw from the same credit balance, so one busy agent can drain credits another team needs.
- Recurring features can use credits as soon as you turn them on, and again each month.
Credit Budget Example: Customer Agent at 500 Conversations
Say a support team on Service Hub Professional expects the Customer Agent to fully resolve 500 text conversations a month. All figures are from HubSpot’s catalog and Service Hub pricing page, as of October 2026:
| Line | Credits | Cost at $0.01 per credit |
|---|---|---|
| 500 resolved conversations × 50 credits | 25,000 | $250 |
| Minus credits included with Professional | (3,000) | ($30) |
| Extra credits to buy each month | 22,000 | $220 |
That $220 is on top of your seat prices, and it assumes no other team uses credits. Only conversations the agent resolves without human help count, so handoffs to a person do not use these credits. For comparison, 100 leads from the Prospecting Agent use 10,000 credits, or $100.
Is HubSpot AI Safe? Data, Privacy and Model Training
HubSpot’s AI safety model relies on:
- Permission-based CRM access
- Account roles and user controls
- Connected-app permissions
- Model cards and third-party AI provider disclosures
- Data minimization
- Human review
- Restrictions on sensitive customer information
- Regular permission audits
What HubSpot itself says (October 2026): its model cards list “Model Zero Data Retention” for its AI providers and note that PII redaction is available for some AI features. We did not find a data-residency statement on HubSpot’s AI trust page, so ask HubSpot directly if you need data kept in a specific region. These are HubSpot’s claims, not our legal assessment.
HubSpot AI vs Alternatives: Quick Decision Guide
If HubSpot is not your CRM yet, or you only need one AI job done, another tool may fit better. Pricing models below are from each vendor’s own page, as of October 2026:
| If you need… | Consider | How its AI is priced |
|---|---|---|
| AI across marketing, sales and service in one CRM you already use | HubSpot (Breeze and Agent Hub) | Seats plus HubSpot Credits; Breeze Assistant included |
| Enterprise agents on a large Salesforce setup | Salesforce Agentforce | Flex Credits ($500 per 100,000 credits), $2 per conversation, or per-user add-ons |
| A support-only AI agent | Intercom Fin or Zendesk AI agents | Fin: $0.99 per resolution; Zendesk: per automated resolution, with some included per agent |
| A low-cost CRM where you bring your own AI model | Zoho CRM (Zia Agents) | Agents free; you pay the model vendor for tokens |
| A simple sales pipeline for a small team | Pipedrive | AI features bundled into plans from $14 per seat (annual) |
| A CRM inside a Microsoft 365 company | Dynamics 365 Sales | Sales agents (preview) included in plans; extra agents use Copilot Credits |
| AI on top of HubSpot data without buying agents | Claude with HubSpot’s connector | Connector free; paid Claude plan required |
Common Mistakes to Avoid When Using HubSpot AI
These six mistakes cost teams the most time and money when they roll out Breeze:
- Turning on every agent at once
- Start with one workflow, measure results, then expand. Activating Prospecting, Customer and Data Agents simultaneously makes it hard to tell what’s actually working.
- Skipping the CRM cleanup step
- Breeze drafts and recommendations are only as accurate as the underlying records; duplicate contacts and stale deal stages produce noticeably weaker output.
- Ignoring credit consumption until the invoice arrives
- Set usage alerts early, especially for the Customer and Prospecting Agents, since their per-outcome credit costs add up quickly at volume.
- Publishing AI drafts without review
- Treat Breeze output as a strong first draft, not a final answer; this applies to outreach emails, blog posts and support responses alike.
- Assuming one plan fits every team
- A Starter-level marketing team and an Enterprise-level sales org will hit very different feature ceilings; check the specific agents included in your tier before committing to a rollout timeline.
HubSpot AI Guide: Final Verdict and Next Steps
HubSpot Breeze is most valuable when a business already uses HubSpot and wants AI grounded in customer, campaign and sales data rather than a disconnected chatbot. Its assistant, specialized agents and embedded features can reduce repetitive work across marketing, sales, service and operations. Advanced automation depends on your plan, clean CRM data and HubSpot Credits. Work out the full monthly cost before you scale up.
Start with one measurable workflow, establish human-review rules, and monitor accuracy and credit usage before enabling additional agents.
HubSpot AI Guide FAQs
Is Breeze Assistant free?
Yes. HubSpot says Breeze Assistant is included with every HubSpot subscription at no additional cost. It answers questions about your CRM data, drafts emails and reports, and helps you prepare for meetings. Agents in Agent Hub are different: most use HubSpot Credits each time they complete a task.
What LLM does HubSpot Breeze use?
HubSpot’s AI model cards list models from OpenAI, Anthropic and Google for Breeze Assistant and the agents, and also Perplexity for some AI features such as AEO (as of October 2026). HubSpot says it uses some of its own models too. The exact models can change, so check HubSpot’s model cards for the current list.
Do unused HubSpot Credits roll over?
No. Unused monthly credits expire at the end of each usage period and do not carry over. Your credit limit resets to its normal level at the start of the next period. Plan agent use across the month, and set alerts so you do not run out early.
What does HubSpot AI Search Grader measure?
It gives a free score out of 100 for how your brand appears in ChatGPT, Perplexity and Gemini answers. The score combines sentiment, presence quality, brand recognition, share of voice and market competition. It is a one-time check. Ongoing tracking is a separate paid product, HubSpot AEO.
Is HubSpot AI included with the free CRM?
Does HubSpot have AI on the free CRM? Partly. HubSpot says Breeze Assistant is included with every HubSpot subscription. Agents are different: they spend HubSpot Credits, and monthly credits only come with paid Starter, Professional or Enterprise plans and paid seats. The free plan also includes limited rule-based chatbots.
Does HubSpot have an AI chatbot?
Yes. The best HubSpot AI chatbot for support is the Customer Agent. It answers text conversations from your knowledge base and uses 50 HubSpot Credits per resolved conversation. Here is when to use each chatbot in HubSpot: rule-based chatflows for simple routing and fixed questions, and the Customer Agent for real customer questions.
