Mike HeapMike is an experienced Product Manager who focuses on all the “non-development” areas of My AskAI, from finance and customer success to product design, copywriting, testing and more.
Table of Contents
- What is HubSpot Breeze AI?
- How easy it is to set up Breeze AI?
- What channels does Breeze AI work in?
- What are the limitations of Breeze AI?
- Content-driven performance (the #1 real-world limiter)
- External knowledge can require workarounds
- Assistant rate limits
- Plan requirements + credits
- Knowledge vault limits and beta surface area
- What knowledge sources can I train Breeze AI on?
- HubSpot-hosted content
- Public URLs / website content
- File uploads (documents)
- CRM data (permissioned)
- What features does Breeze AI have?
- The Customer Agent (automated tier-1 support)
- Handoff and routing controls
- Personality / tone configuration
- Actions (API calls triggered by phrases)
- Analytics and monitoring
- Context tooling via knowledge vaults
- How do I improve Breeze AI responses?
- What resolution rate can I expect from Breeze AI?
- What AI model does Breeze AI use?
- What languages does Breeze AI work in?
- How secure is Breeze AI?
- Who is using Breeze AI?
- How much does Breeze AI cost?
- Does Breeze AI have a free trial?
- Is Breeze AI worth it?
- FAQs
Last Updated
Created time
Mar 3, 2026 02:15 PM
Title length (<60)
Author
Ecomm?
Tags
Image
Publish date
Mar 3, 2026
Slug
hubspot-breeze-ai-agent-complete-guide-2026
Featured
Featured
Type
Article
Ready to Publish
Ready to Publish
I know why you’re here…
You’re here because of one of three reasons:
- You got the Breeze pitch, saw the credits math, and immediately started doing “cost per conversation” back-of-napkin calculations.
- You tried to set it up and realized your support “knowledge” does not, in fact, live neatly inside HubSpot Knowledge Base articles.
- You ran it in production, and it was… fine. But not “we can finally breathe” fine.
This post is the time-saver version: what Breeze is, what it actually does inside Service Hub, what it can (and can’t) learn from, how to roll it out safely, what resolution rates are realistic, and how the credits model changes the economics.
What is HubSpot Breeze AI?
TL;DR: Breeze is HubSpot’s umbrella AI platform across marketing, sales, and service, made up of a conversational assistant plus task-specific agents like the Customer Agent for support. It launched as a major product push in September 2024.
Breeze is HubSpot’s “AI layer” across the whole CRM platform, not just a support bot. HubSpot positions it as a unified system spanning marketing, sales, and service, including a chat-style assistant and multiple specialized agents you can configure and deploy inside HubSpot tools (Service Hub included) via Breeze and the Breeze AI Customer Agent.
In practical terms, Breeze includes:
- A conversational assistant (“Breeze Assistant”) that helps users generate, summarize, and answer questions inside HubSpot, including with connected apps via Breeze Assistant connected apps.
- “Breeze Agents” that are more like role-based teammates. For service, the important one is the Customer Agent, which can reply to customers across HubSpot-supported channels and hand off when needed.
- Admin tooling for configuring context and deployments, including managing Breeze context with knowledge vaults.
HubSpot made Breeze a headline release in September 2024. And while HubSpot talks a lot about sales qualification and lead gen use cases for agents, this guide stays focused on the service side: the Customer Agent inside Service Hub.
How easy it is to set up Breeze AI?
TL;DR: Setup is designed to be no-code and fast—HubSpot suggests most teams can get a Customer Agent running in under 15 minutes—assuming your channels and knowledge sources are already in place.
HubSpot’s Customer Agent setup is genuinely straightforward, as long as your basics are already true (you have channels connected, and you have content worth “reading”).
The core flow is documented in set up and configure the customer agent and create a customer agent. It’s essentially: name the agent, pick a role/personality, connect knowledge, assign to channels, test, then go live.
HubSpot (and partners writing about it) describe this as quick—“most teams can do this in under 15 minutes” is a commonly-cited expectation in writeups like this Breeze Customer Agent overview.
The catch is that “set up the agent” and “deploy something that won’t embarrass you” are not the same thing.
You’ll usually spend more time on:
- curating what content the agent should use
- setting handoff rules and schedules
- and running enough test questions that you trust it
HubSpot supports testing before going live, and you’ll want to use it.
Breeze performance is content-driven: if the KB is thin, outdated, or inconsistent, you’ll see that immediately in escalations and “unanswered questions.”
My AskAI is a direct swap for HubSpot Breeze Customer Agent, with an approved app in the HubSpot marketplace, entirely no-code, you can set it up in under 10 minutes and save 10-12x the cost.
What channels does Breeze AI work in?
TL;DR: Breeze Customer Agent can run wherever HubSpot’s inbox/conversations run—live chat, email, WhatsApp, and Facebook Messenger—with voice/phone called out as a beta channel in HubSpot materials.
Breeze Customer Agent is built to respond on the core channels HubSpot already supports through Conversations/Inbox.
HubSpot explicitly markets “consistent support across chat, WhatsApp, Facebook, email, and voice 24/7” on the Breeze AI Customer Agent page.
In channel terms, you’ll commonly see it used with:
- Live chat via HubSpot Chatflows (and note: to use chat on an external website, HubSpot calls out that you need the tracking code installed, per customer agent setup prerequisites)
- Email (via the HubSpot inbox / helpdesk routing)
- Facebook Messenger
- Voice/phone is referenced as a capability in HubSpot’s messaging, and appears as a beta in channel lists described in ecosystem writeups and HubSpot materials linked from the same product pages
So if your support operation lives inside Service Hub’s omnichannel inbox, Breeze is meant to “just fit” without duct tape.
My AskAI works in all HubSpot channels.
What are the limitations of Breeze AI?
TL;DR: The big constraints are (1) content quality and coverage, (2) usage/credit economics, (3) some rate limits on the Assistant, and (4) source/connectivity limitations when your knowledge lives outside HubSpot.
Breeze is powerful, but it’s not magic. The limitations are mostly predictable—and if you plan around them, you’ll have a much better rollout.
The practical constraints called out across HubSpot docs and summaries include:
Content-driven performance (the #1 real-world limiter)
Breeze can only answer from what it can reference.
If content is missing or unclear, you get gaps, escalations, or low-confidence answers. HubSpot and partners emphasize content formatting and relevance as a best practice in guides like improving agent performance and content structure.
External knowledge can require workarounds
Out of the box, Breeze’s “cleanest” knowledge source is HubSpot-hosted content (especially HubSpot Knowledge Base articles).
If your KB lives somewhere else (Notion, Confluence, Zendesk, etc.), you may be relying on URL imports or file uploads instead of a first-class native connector.
This limitation is also described bluntly in public commentary like this HubSpot community discussion on external KBs.
Assistant rate limits
HubSpot documents usage limits for the assistant: Breeze Assistant is described with request limits (commonly cited as around 30 requests per minute and 1,000 per day). This is more relevant to internal users hammering the assistant than the Customer Agent, but it matters for ops planning if teams want to use Breeze across departments.
Plan requirements + credits
Breeze Agents (including the Customer Agent) are positioned as Pro/Enterprise capabilities, and usage is metered via HubSpot Credits as described across HubSpot pricing and credit pages like HubSpot credits pricing. Even if setup is easy, the ongoing cost model is a real constraint.
Knowledge vault limits and beta surface area
Knowledge Vaults exist, but there are platform limits (for example, HubSpot documents vault management in knowledge vaults, including caps like a maximum number of vaults). Also, some functionality is explicitly beta (like parts of Breeze Studio), so you should expect iteration.
My AskAI charges based on ticket volume, so you can more easily predict your usage and keep your costs stable as your resolution increases. You can also connect your My AskAI to more knowledge sources for direct replies, including Google Drive, Notion, OneDrive and more.
What knowledge sources can I train Breeze AI on?
TL;DR: Breeze can use HubSpot content (KB, website/blog pages), uploaded files, public URLs, and CRM data (permissioned), plus “knowledge vaults” to bundle context; it’s not designed as a native connector hub for third-party knowledge tools like Notion or Google Drive.
Breeze isn’t “trained” in the classic ML sense; it’s configured with sources it can reference at answer time.
In practice, what matters is: can it reliably retrieve the right chunk of truth when a customer asks a messy question?
Based on HubSpot documentation and Breeze context tooling, the core knowledge inputs are:
HubSpot-hosted content
This is the most natural fit. Breeze can use HubSpot knowledge base articles and other HubSpot content assets you already manage in your portal. This is reflected in how the Customer Agent is configured in manage your customer agent and in broader “what Breeze is” documentation like understand Breeze.
Public URLs / website content
Breeze can ingest from websites/URLs, with options like importing related URLs, as described in guides linked through HubSpot’s agent setup and partner writeups like this Customer Agent setup overview.
File uploads (documents)
Breeze supports using uploaded files (common formats referenced include PDFs and office docs), and HubSpot’s “context management” is centered around knowledge vaults where you can bundle files and CRM segments.
CRM data (permissioned)
Breeze can use CRM context when enabled, including allowing the agent to access and update CRM data as described in allow the customer agent to access and update CRM data.
This is a meaningful advantage if your support workflows depend on customer/account context that already lives in HubSpot.
What Breeze is not positioned as (today) is a “connect to anything” knowledge platform.
For example, when compared directly against third-party tools that market broad knowledge integrations, Breeze is described as limited primarily to HubSpot KB, websites, and static uploads (meaning if a file changes, you may need to re-upload it) in managing the Customer Agent’s knowledge sources.
And it’s publicly acknowledged that external KBs can require workarounds (again, see the external KB limitation discussed in this public HubSpot thread).
My AskAI can create tasks, actions and connect to any API to improve responses. You can also connect your My AskAI to more knowledge sources for direct replies, including Google Drive, Notion, OneDrive and more.
What features does Breeze AI have?
TL;DR: Breeze combines the Customer Agent for automated support with the Breeze Assistant for internal productivity, plus configurable actions, analytics, and context management via knowledge vaults and agent settings.
Breeze’s feature set is broad because it’s not a single feature—it’s a platform across the CRM.
For service teams, the relevant capabilities cluster into a few buckets:
The Customer Agent (automated tier-1 support)
The core job is straightforward: handle routine questions automatically and hand off when it can’t. HubSpot describes this in the Breeze AI Customer Agent product page and the operational setup docs like create a customer agent.
Handoff and routing controls
Breeze supports handoff mechanics inside chat experiences (including handoff messaging and rule configuration during setup), as part of the standard Customer Agent configuration flow described in set up and configure the customer agent.
Personality / tone configuration
Breeze supports defining a role and tone/personality during setup, again part of the customer agent configuration flow in create a customer agent.
Actions (API calls triggered by phrases)
Breeze can be configured with “actions” using trigger phrases and API calls, as referenced in HubSpot’s customer agent tooling and linked topics inside manage your customer agent. This is where Breeze starts to move from “answers questions” to “does something,” though it’s still bounded by what you’ve configured.
Analytics and monitoring
HubSpot provides performance dashboards that report on things like conversations resolved/deflected and “unanswered questions.” These are shown and discussed in the Customer Agent materials and in partner summaries like this Breeze Customer Agent reporting overview.
Context tooling via knowledge vaults
Knowledge Vaults let you package context (files, CRM segments, etc.) in a manageable way described in manage Breeze context with knowledge vaults. The point is to keep context organized and reusable across agents/assistants.
And then, beyond service, Breeze also includes other agents (prospecting, content/blog, data hygiene, etc.) under the broader Breeze platform. That matters if you’re trying to justify credits spend across multiple departments, not just support.
My AskAI also has these features, and in addition, can reply in internal as well as direct modes and has a self-learning feature to automatically improve AI knowledge.
How do I improve Breeze AI responses?
TL;DR: You improve Breeze primarily by improving and curating its source content, then using the unanswered questions/knowledge gaps reporting to iteratively close gaps and retest; handoff rules and feedback loops help prevent low-confidence guessing.
If you treat Breeze as “set it and forget it,” you’ll get exactly what you deserve: a bot that escalates constantly or confidently answers the wrong thing.
Breeze improvement is mostly an ops loop:
- Start with clean, structured content
HubSpot and partner materials emphasize that short, clear sections with headings and bullet points tend to work better than giant walls of text. You see this guidance in resources like this guide to setting up and optimizing the Customer Agent.
- Use customer language, not internal language
If customers ask “Where’s my order?” but your article is titled “Fulfillment tracking protocol,” don’t be surprised when the agent misses. The same optimization guide above emphasizes matching common phrasing patterns.
- Monitor performance dashboards and unanswered questions
Breeze includes dashboards for resolution/deflection and a view that groups unanswered questions by topic, which you can use to decide what content to add next (shown in HubSpot UI screenshots and described in resources like manage your customer agent).
- Add missing answers, then test again
HubSpot supports saving and running test questions, but the key is doing it repeatedly as you change content. This is the difference between “we turned it on” and “it got better.”
- Configure handoff rules conservatively
Handoff rules are not a failure; they’re a safety mechanism. Properly tuned handoff prevents Breeze from guessing in edge cases, and you can configure these during setup via create a customer agent.
- Keep vaults and sources curated
Knowledge Vaults are useful, but “add everything” is not a strategy. HubSpot documents how to manage and scope these in knowledge vault management.
The meta point: Breeze improves when your knowledge improves. The agent is a mirror held up to your documentation discipline.
My AskAI allows you to inspect responses to better understand answers, has QA and testing features and also self-learns to fix knowledge gaps.
What resolution rate can I expect from Breeze AI?
TL;DR: HubSpot markets 65%+ resolution with top teams reaching ~90%, and its ROI modeling often uses ~68% as a baseline; real-world outcomes vary widely based on content quality and query complexity.
HubSpot’s public positioning sets expectations that “good setups deflect a lot.”
Specifically, HubSpot states that AI resolutions are 65%+ with some teams hitting 90%. Partner calculators and ROI models commonly use benchmarks around the high 60s (for example, a default benchmark of ~68% appears in ROI modeling referenced in resources like this Breeze credits cost calculator).
Case studies show variability. HubSpot publishes customer stories here: Breeze customer success stories. Individual stories include outcomes like resolving a meaningful share of chat inquiries, but there’s no single universal “Breeze average” publicly presented as a platform-wide aggregate metric.
A realistic expectation-setting approach:
- If you start with a decent KB and fairly repetitive questions, you can aim for the 50–70% band relatively quickly.
- If you iterate aggressively on knowledge gaps and the agent’s sources are high quality, you can push higher.
- If your support is highly technical, highly account-specific, or your knowledge is scattered/outdated, you should expect more handoffs until you fix the inputs.
My AskAI has resolved over 1.1m tickets to date with an average resolution rate of 72%.
What AI model does Breeze AI use?
b: HubSpot publishes model cards showing Breeze uses OpenAI GPT-4-class models (including GPT-4o and GPT-4.1 variants) for language and Stability AI models for images, with additional models (including Anthropic Claude) used in some embedded AI features.
HubSpot is unusually transparent here—there are published model cards.
According to HubSpot’s AI Model Cards:
- Breeze Agents use OpenAI GPT models (including GPT-4.1, GPT-4.1 Mini, and GPT-4o variants) for language tasks, and Stability AI models for image generation where relevant.
- Breeze Assistant uses OpenAI GPT-4o and GPT-4o Mini (plus Stability AI for images).
- Some embedded AI features also use additional models, including Anthropic’s Claude (as listed in those model cards).
HubSpot also states “Model Zero Data Retention” in the model card documentation, meaning customer data isn’t used to train the underlying models, as described in the same AI model cards.
My AskAI use a suite of different models, namely from OpenAI and Google. We are also constantly running A/B tests and benchmarking to determine whether better quality, faster models can be used.
What languages does Breeze AI work in?
TL;DR: Breeze Customer Agent and Breeze Assistant are available in all HubSpot-supported languages, but quality depends on having source content in the target language and testing performance in your real tickets.
HubSpot states that Breeze is available across HubSpot-supported languages, as referenced in resources like the Breeze AI customer agent page.
The operational reality is straightforward: multilingual output depends heavily on multilingual input.
If your KB is English-only, don’t expect magical accuracy in German. Test it in the languages you care about before you roll it out broadly.
My AskAI is multilingual in 95 different languages.
How secure is Breeze AI?
TL;DR: Breeze inherits HubSpot’s security model, including permissioning (it can only access what the user can access), masking/excluding sensitive fields, and region-based processing, with compliance coverage including GDPR, HIPAA, and CCPA.
Security is one area where HubSpot tends to be strong, and Breeze is explicitly built within HubSpot’s existing trust framework.
Key security/privacy behavior is explained in detail in Data Privacy: What HubSpot Breeze Can See in Your CRM, including:
- Permission-aware access: Breeze respects HubSpot user permissions and doesn’t “see” data a user cannot see.
- Sensitive data handling: HubSpot describes masking/excluding highly sensitive information (PHI/PII categories) from AI processing.
- Regional processing: HubSpot describes data processing staying in-region (EU stays EU, etc.).
- Admin controls and auditability: admins can control AI access and features.
HubSpot also maintains a broader trust portal with certifications and controls at HubSpot trust and security, and the AI-specific model behavior is documented in HubSpot’s AI model cards.
If you’re in a regulated industry, Breeze being “inside HubSpot” is generally an advantage versus bolting on a separate AI system—because it keeps data flows within HubSpot’s environment, with HubSpot’s controls.
My AskAI is SOC-2 Type II certified and GDPR compliant.
Who is using Breeze AI?
TL;DR: Breeze is used by thousands of HubSpot customers, and HubSpot highlights case studies across industries (e-commerce, SaaS, education, and more) with reported efficiency and support outcomes.
HubSpot claims broad adoption (“thousands of businesses trust Breeze”), and publishes named customer stories and outcomes on the Breeze AI case studies page.
Examples referenced in public Breeze stories and summaries include companies like Nutribees, Agicap, Sandler, Kaplan, and Stübben (among others) appearing across HubSpot’s Breeze marketing and case studies ecosystem.
The most useful way to read these: look for companies whose support motion resembles yours (channels, ticket types, documentation maturity), not just impressive percentage claims.
My AskAI is used by over 200 businesses around the world, with ticket volumes from a few hundred per month up to 100k per month. You can read our recent case studies here.
How much does Breeze AI cost?
TL;DR: Breeze Customer Agent usage is metered via HubSpot Credits, with 100 credits per conversation and example pricing of $50 for 5,000 credits (≈$1 per conversation), and it requires Pro/Enterprise hub plans for full agent capabilities.
Breeze Customer Agent runs on HubSpot’s credits system, documented at HubSpot credits pricing and in related calculators like HubSpot’s own Breeze Customer Agent ROI calculator.
The concrete usage unit (for the support agent) is:
- 100 credits per conversation the agent is involved in, per the Customer Agent marketplace listing.
And HubSpot’s ROI calculator example pricing commonly cited is:
- $50 for 5,000 credits, from the Breeze ROI calculator.
If you do the math using that example:
- 5,000 credits / 100 credits per conversation = 50 conversations
- $50 / 50 conversations = $1 per conversation
So if you have 10,000 conversations per month where the agent participates, that’s roughly $10,000/month in credits on that example pricing.
One thing to note is the model is “per conversation,” not “per successful resolution.” If the agent escalates, you still spent the credits for that conversation.
Finally, there’s also the plan requirement: Breeze Agents are positioned as Pro/Enterprise functionality (with free/Starter having limited AI features), as summarized in the broader Breeze packaging on Meet Breeze.
My AskAI plans start from $199/mo, with additional tickets costing from $0.10 per ticket. This works out around 10-12x cheaper than Breeze. You can compare your ROI for Breeze AI vs My AskAI here.
Does Breeze AI have a free trial?
TL;DR: HubSpot offers free CRM access (with some AI capabilities like Breeze Assistant), but full Breeze Agent capabilities (like Customer Agent automation) are tied to paid HubSpot plans and credit usage.
HubSpot’s positioning is that some AI functionality exists in the free CRM experience, while full agent deployment requires a Pro/Enterprise plan.
This packaging distinction is described at a high level on Meet Breeze and reinforced by the requirement framing across Breeze/Customer Agent setup and credits.
Practically: you can explore Breeze in a HubSpot portal and trial elements, but if you want the Customer Agent running at scale, you’re in paid-plan + credits land.
My AskAI has a 30 day, unlimited free trial. Test every feature to your heart’s content and don’t pay a dime. No credit card required.
Is Breeze AI worth it?
TL;DR: Breeze is worth it when you want a native, permission-aware AI agent tightly integrated with HubSpot CRM and channels, and you can justify ongoing per-conversation credit costs; it’s harder to justify if your knowledge lives outside HubSpot or you’re cost-sensitive on high ticket volumes.
Breeze is a good product in the way “built into the platform” products are usually good.
Breeze is compelling when:
- Your support operation already lives in HubSpot Conversations/Help Desk.
- Your knowledge base is already in HubSpot Knowledge Base (or you’re willing to move/reshape it).
- You want something that inherits HubSpot’s permission model and governance, which is a real advantage described in Breeze CRM visibility and permissions behavior.
- You want AI across departments (marketing + sales + service) and can amortize credits spend across multiple teams.
Breeze is harder to justify when:
- You have high conversation volume and the “$1 per conversation” credits example (from Breeze ROI calculator pricing) makes the ROI math painful.
- Your “truth” lives outside HubSpot (Notion, Google Drive, Confluence, external helpdesks), and you don’t want to play workaround games, which is a known limitation acknowledged publicly in discussions like this external KB note.
- You want more granular improvement tooling than basic unanswered question grouping and manual test runs (Breeze has reporting and an unanswered questions view, but the improvement loop is largely “you fix the content and test again” per Customer Agent management).
And yes: there are third-party alternatives that plug into HubSpot.
One example is the My AskAI HubSpot marketplace app with more detail at My AskAI’s HubSpot integration page. If your core goal is “reduce cost versus native helpdesk AI,” the existence of alternatives matters—especially given Breeze’s credits economics.
The sane rollout path (if you’re unsure) is a pilot:
- pick a limited set of FAQ-style topics,
- deploy in one channel first,
- monitor deflection/resolution and unanswered questions,
- tighten knowledge sources and handoff rules,
- then expand.
That approach aligns with how HubSpot frames iteration via analytics and knowledge gaps in Customer Agent management.
If you want a full comparison of My AskAI against Breeze AI, read one here.
FAQs
Is Breeze AI the same thing as the Breeze Customer Agent?
No. Breeze is the umbrella AI platform across HubSpot, while the Customer Agent is a specific Breeze Agent focused on automated customer support, described on the Breeze AI Customer Agent product page.
What’s the difference between Breeze Assistant and Breeze Agents?
Breeze Assistant is a conversational helper inside HubSpot for internal users (generating/summarizing/searching), while Breeze Agents are task-focused “teammates” like the Customer Agent for service workflows, described in understand Breeze.
Which HubSpot plans do I need to use the Breeze Customer Agent?
Breeze Agents are positioned as Professional/Enterprise features in HubSpot’s Breeze packaging, summarized on Meet Breeze, and usage is metered with credits per HubSpot credits pricing.
How does Breeze Customer Agent pricing work?
The Customer Agent consumes HubSpot Credits, with 100 credits per conversation per the Customer Agent listing. HubSpot’s ROI calculator example pricing shows $50 for 5,000 credits on the Breeze ROI calculator, which works out to about $1 per conversation using that example.
What channels can Breeze Customer Agent support?
HubSpot markets support across chat, WhatsApp, Facebook, email, and voice on the Breeze Customer Agent page. Channel availability depends on what you’ve connected in HubSpot Conversations and the prerequisites in customer agent setup documentation.
Can Breeze Customer Agent use my CRM data to personalize answers?
Yes, if configured. HubSpot documents allowing the Customer Agent to access and update CRM data in allow the customer agent to access and update CRM data, and Breeze respects permissions as explained in what Breeze can see in your CRM.
What knowledge sources can Breeze use?
Breeze can use HubSpot-hosted content, public URLs/websites, file uploads, and bundled context via knowledge vaults. Customer Agent knowledge management is described in manage your customer agent.
Does Breeze have limitations if my knowledge base is in Notion/Confluence/Zendesk?
Breeze works best with HubSpot-hosted KB content, and external KBs can require workarounds. This limitation is also acknowledged publicly in discussion like this note about external knowledge bases.
How do I know what Breeze failed to answer?
Breeze includes performance analytics and an “unanswered questions” view that groups missed questions by topic, referenced in the Customer Agent management flow at manage your customer agent and discussed in optimization writeups like this Breeze Customer Agent guide.
What resolution rates does HubSpot claim for Breeze Customer Agent?
HubSpot states that AI resolutions are 65%+ with some teams hitting 90% on the Breeze Customer Agent page.
What AI models power Breeze?
HubSpot publishes model usage in HubSpot’s AI Model Cards, describing OpenAI GPT-4-class models (including GPT-4o and GPT-4.1 variants) for language tasks and Stability AI models for image generation, with additional models (including Anthropic Claude) used in some embedded features.
How does Breeze handle security and permissions?
Breeze enforces HubSpot permissions and follows HubSpot’s privacy controls. A detailed explanation is provided in what Breeze can see in your CRM, and HubSpot’s overall security posture is documented at HubSpot trust and security.
Is there a free version of Breeze?
HubSpot offers free CRM access and some AI capabilities, while full Breeze Agent deployment is positioned as part of paid plans and credit usage, summarized on Meet Breeze and governed by HubSpot credits pricing.
Where can I see real examples of companies using Breeze?
HubSpot publishes customer stories on the Breeze AI case studies page.
