Updated October 8, 2026.

What is HubSpot Breeze?

Breeze is HubSpot’s AI, connecting conversational assistance and AI capabilities with the customer platform. Breeze Assistant helps users work with CRM context. Agent Hub, formerly Breeze Agents, provides a home for agents, while Agent Builder, formerly Breeze Studio, supports creating and customizing them. Availability and usage costs vary by capability.

Source: HubSpot’s Breeze Assistant overview and product FAQs.

For businesses using HubSpot, the opportunity is to make everyday work more useful and consistent: preparing for customer conversations, researching accounts, drafting content, and managing repeatable processes.

But choosing where to start requires understanding which tool does what—and what your team needs to configure, review, and measure.

This guide explains the main components of HubSpot AI and how to approach adoption with a clear business purpose.

How Breeze Assistant, Agent Hub, and Agent Builder fit together

Older articles and screenshots may use names that have since changed. Here is the current relationship:

Agent Hub was previously called Breeze Agents. Agent Builder was previously called Breeze Studio.

Sources: Breeze Assistant FAQs and Agent Builder overview.

For a focused explanation of the naming changes, read Agent Hub vs. Breeze: What’s the Difference?.

Breeze Assistant: help while you work

Breeze Assistant supports CRM-grounded questions, drafting, meeting preparation, pipeline analysis, and building agents or workflows through natural-language instructions. HubSpot includes the Assistant with subscriptions at no additional cost.

Source: HubSpot Breeze Assistant.

A useful starting point is a task your team already understands well.

For example, a seller preparing for an account meeting could ask for a summary of relevant customer information, then check the result against the underlying records. A marketer could use an initial draft to accelerate writing while retaining responsibility for positioning, accuracy, and the final message.

Treat prompts as working instructions. Explain the audience, the desired output, and the information that matters. Review whether the response actually answers the question before using it in a customer interaction or business decision.

Agent Hub: managing agents across the business

When multiple teams introduce agents, ownership becomes a practical concern.

Who configured the agent? What process does it support? Who investigates an unexpected result? Who decides whether it should expand to additional users or workflows?

Use your agent management process to make those responsibilities visible. Maintain a clear purpose, owner, and success measure for each deployment.

For the product walkthrough, read What Is Agent Hub? A Complete Guide and How Agent Hub Works.

Agent Builder: adapting AI to your process

Agent Builder supports custom agents using your prompts, knowledge, and CRM data. Custom runs consume credits based on action units, and the work performed can vary between runs.

Source: HubSpot Agent Builder.

Before building a custom agent, write down the process it should support:

• What starts the task?
• What information should it use?
• What output should it produce?
• What actions are acceptable?
• When should a person review or take over?

For example, an account-research process needs a definition of a useful finding. “Research this company” leaves considerable room for interpretation. A more specific instruction could identify the business signals to investigate, the format for results, and how to handle missing evidence.

Customization becomes easier to evaluate when the desired behavior is explicit.

Understanding the individual agents

HubSpot’s agent capabilities cover different kinds of work. Customer Agent supports automated support responses and resolution. Prospecting Agent supports account research and personalized outreach. Data Agent supports customer research and insights.

Source: HubSpot’s agent overview.

Choose the detailed resource that matches your intended use:

Check the current requirements for the individual capability before planning deployment. Avoid assuming that every agent supports the same inputs, actions, channels, or approval settings.

Why CRM data quality matters

Connecting AI to CRM information gives you a reason to examine how reliable that information is.

A customer summary may be incomplete when important conversations were never recorded. Research can become harder to interpret when duplicate companies carry conflicting information. An automated process can produce the wrong outcome when its routing criteria or lifecycle definitions are unclear.

Before a pilot, review the data that the chosen workflow actually needs. Prioritize the records, properties, and source content involved in that use case rather than postponing adoption until the entire database is perfect.

For a practical readiness review, read Do I Need Clean CRM Data Before Using Agent Hub?.

Permissions, privacy, and human oversight

Review access and data handling for each feature you enable. HubSpot documents tool-specific considerations for Sensitive Data, including additional restrictions for Highly Sensitive Data.

Source: Understand how Sensitive Data is used in HubSpot tools.

Your implementation review should answer:

• Which records and content are appropriate inputs?
• Who

Related Resources
Data Privacy Deep Dive: What HubSpot Breeze Can (and Can’t) See in Your CRM
Unified Intelligence: How Breeze Uses Smart CRM Data to Power Every Interaction
The Human + AI Partnership: How Teams Should Really Use Breeze
Everyday Use Cases: How Marketing, Sales, and Service Teams Use HubSpot Breeze
How to Get Started with HubSpot Breeze: Setup, Training, and Team Adoption
Measuring the Impact: How to Track ROI from HubSpot Breeze
The Future of HubSpot AI: What’s Next for Breeze Agents and Smart CRM
How Breeze Changes the Way Businesses Think About Growth