What HubSpot's Breeze actually is
HubSpot has consolidated its AI under a single brand called Breeze, organized into three areas: Breeze Assistant, Breeze Agents, and Breeze Intelligence. Rather than bolting AI onto individual tools, HubSpot positions Breeze as a layer that runs across its Smart CRM, drawing on the customer data already in the platform.
If you're evaluating "HubSpot AI features," it helps to know the naming history. The conversational component, Breeze Assistant, was previously called Breeze Copilot, and before that ChatSpot. The Breeze brand itself is current; the Copilot-to-Assistant rename is the main one to keep straight when you read older documentation or third-party reviews.
Breeze Assistant: the conversational layer
Breeze Assistant is HubSpot's in-app conversational AI. It's where users ask questions in natural language, summarize CRM records, generate content, and get answers to product questions. Because it's grounded in your CRM data, the idea is that responses reflect your actual contacts, deals, and records rather than generic output.
HubSpot also lets teams build custom assistants on top of Breeze, so the same conversational interface can be shaped around specific workflows or knowledge. For most buyers, the Assistant is the lowest-risk entry point: it drafts, summarizes, and answers, but a person stays in the loop to review what it produces before anything ships.
Breeze Agents: AI that takes action
Breeze Agents are autonomous, task-specific agents that go beyond answering questions to performing work. Three are generally available:
- Customer Agent — resolves inbound support inquiries automatically with CRM-grounded responses, acting as an always-on first-line support specialist across channels.
- Prospecting Agent — surfaces in-market accounts based on buying signals, builds contact lists, and equips reps with personalized outreach.
- Data Agent — answers questions about customers by analyzing CRM data, conversations, documents, and the web.
HubSpot lists additional agents in beta, including a Customer Health Agent (evaluates account health, prioritizes at-risk customers, and drafts outreach), a Company Research Agent (pulls insights from a target company's website, recent news, and CRM records), and a Closing Agent (surfaces deal blockers and drives buyer engagement). Agents can be configured with Tools that define the actions an agent is allowed to perform and Knowledge vaults that supply reference information.
Breeze Intelligence: enrichment and buyer intent
Breeze Intelligence is a data enrichment and buyer-intent solution built into the Smart CRM. It enriches company and contact records using firmographic, demographic, and technographic attributes drawn from a dataset of company and buyer profiles, so records fill in with less manual research.
On the intent side, marketers can define a target market, identify companies showing intent signals (including those visiting your website), and add high-fit companies to the CRM. Breeze Intelligence also offers form shortening, which dynamically hides form fields when the data is already known — shorter forms for visitors, fewer gaps in your records.
How to judge whether Breeze stays inside your guardrails
Once AI moves from drafting to acting — resolving a support ticket, sending outreach, writing to a record — the right question shifts from "is it smart?" to "can I see and undo what it did?" That's the lens we apply to any tool that acts on your account: bounded autonomy. We frame every action as Trigger, Action, Impact — what set the agent off, what it did, and what changed as a result. If you can't reconstruct all three for a given action, the automation is a black box, regardless of vendor.
Breeze gives you real controls to work with here. The Tools mechanism defines what an agent is permitted to do, and Knowledge vaults scope what it references — both are levers you set before an agent acts. As you evaluate, push on the specifics: Does the Customer Agent log its reasoning before it replies to a customer? Can you require human review on outbound steps the Prospecting Agent prepares? Is every enriched field or write-back attributable and reversible? Our take on guardrail-driven automation is to start with read-only or review-required access, watch the logs, and widen autonomy only where the impact is visible and reversible. That's our evaluation framework, not a verdict on HubSpot — the point is to apply the same test to every vendor on your shortlist.
Last reviewed June 2026. HubSpot updates its AI capabilities frequently — confirm current specifics on the vendor's own site.
Before you turn on a single Breeze Agent, know where your guardrails should go
The Readiness Score shows you which AI actions are safe to automate on your stack today and which need a human in the loop first. 4 minutes, no login.
Get your free Readiness Score →Keep reading
- AI agents in your CRM — what changes when agents act on your records, and how to keep control
- The bounded-autonomy buyer's guide — the full checklist for judging any tool that acts on your account
- Salesforce AI features explained — compare HubSpot's Breeze against another CRM-native AI suite