A practical guide to HubSpot Breeze Agent inputs and outputs

Kenneth Pangan
Written by

Kenneth Pangan

Stanley Nicholas
Reviewed by

Stanley Nicholas

Last edited October 16, 2025

Expert Verified

Everyone's talking about how AI agents are going to change the way we work, and HubSpot is right in the mix with its Breeze Agents. They're pitched as "digital coworkers" that can automate tasks right inside your CRM. But before you jump in, it’s worth looking past the hype. To figure out if they'll actually work for your team, you need to get a handle on two simple things: what you have to put in to get them started, and what you actually get out.

This guide is a straight-up look at what it takes to get Breeze Agents running. We’ll dig into their use cases, pricing, and some of the tricky parts you might not see at first glance. By the end, you’ll have a much clearer idea of whether they’re the right move for you, or if a more flexible tool would make more sense.

What are HubSpot Breeze agents?

So, what exactly are HubSpot Breeze Agents? Think of them as AI tools that live inside HubSpot and are built to handle whole jobs, not just little tasks. A simple AI assistant might help you draft an email, but a Breeze Agent is designed to take a task, like writing a blog post or solving a support ticket, and see it through from start to finish on its own.

A look at the HubSpot AI Agent interface, demonstrating the core concept of Breeze Agent inputs and outputs.::
A look at the HubSpot AI Agent interface, demonstrating the core concept of Breeze Agent inputs and outputs.

They're just one piece of the larger Breeze AI puzzle, which also includes:

  • Breeze Copilot: This is more of an on-demand assistant. It works alongside you to help with things like summarizing calls or writing content, but it waits for your command.

  • Breeze Intelligence: This is the data layer working behind the scenes, pulling in company info and buyer intent signals to give the AI more context.

The Agents are the real workhorses here, operating in the background once you've set them up. And understanding the Breeze Agent Inputs and Outputs is the secret to getting any real value from them.

The inputs: What you need to get started

Any AI is only as good as the information it's given, and Breeze Agents are no different. They are wired directly into the HubSpot ecosystem, meaning they pretty much only "eat" the data and settings you have inside that one platform.

HubSpot data

Breeze Agents get almost all their context from your HubSpot account. This means a few things are critical:

  • CRM Data: Your contact, company, and deal records are the foundation for personalizing sales emails and understanding customer history. If your CRM data is a mess, the agent's work will be, too.

  • Knowledge Base: The Customer Agent leans entirely on your official HubSpot knowledge base to answer questions. If that KB is gathering dust or missing key articles, the agent won't be much help.

  • Past Content & Performance: To learn your brand voice, the Content and Social Agents look at your old blog posts, landing pages, and social media activity.

Screenshot of a HubSpot Knowledge Base, a critical element for Breeze Agent inputs and outputs.::
Screenshot of a HubSpot Knowledge Base, a critical element for Breeze Agent inputs and outputs.

On one hand, this tight integration is great. On the other, it creates a bit of a "walled garden." If your team’s important info is spread out across tools like Google Docs or Confluence, Breeze Agents simply can't see it. The only way to fix this is to manually move everything into HubSpot, which can be a huge, time-sucking project.

Pro Tip
This is actually a common headache, and tools like eesel AI are built to fix it. Instead of making you move all your data, eesel just connects to the tools you already use (your help desk, wikis, docs, etc.) in a few minutes. It gives the AI a complete picture of your company knowledge without you having to migrate a single file.

The setup: Telling the agent what to do

Getting a Breeze Agent running isn't a one-click affair. You have to spend some time in HubSpot's workflow or agent studio giving it clear instructions. These "prompts" tell the agent its goals, what tone to use, and when to pass a task to a human. For instance, you could tell the Customer Agent to handle all password reset questions but to immediately escalate anything with the word "refund." The interface is user-friendly, but writing good prompts that account for every possible scenario is a skill that takes time and tweaking to get right.

The HubSpot Workflows interface, where users configure the Breeze Agent inputs and outputs.::
The HubSpot Workflows interface, where users configure the Breeze Agent inputs and outputs.

The outputs: What you actually get

Once you've fed the agent the right information, it starts producing results. These can be anything from a finished piece of content to an action taken in your CRM. Here’s a look at what that actually looks like for different teams.

Content and communication

This is what you'll see most often: agents creating text and media for you.

  • Content Agent: It can write entire blog posts, landing page copy, and case studies. It's a fantastic starting point if you have writer's block, but you'll absolutely need a human to edit the draft, check the facts, and add your brand's unique voice.

  • Social Media Agent: This one will create posts tailored for platforms like LinkedIn or X. It can even suggest a content schedule based on what's worked before, but it's not going to jump on a breaking trend or join in a nuanced online conversation.

  • Prospecting Agent: It drafts personalized outreach emails using the info in your CRM. This definitely saves time, but the "personalization" is only as good as the data in HubSpot, which might be missing the real-world business context you'd naturally include.

An example of content generation, a key feature of the Breeze Agent inputs and outputs.::
An example of content generation, a key feature of the Breeze Agent inputs and outputs.

Automated actions

Beyond writing, Breeze Agents can also take action inside HubSpot, which is where they can really save you time.

  • Customer Agent: Its main job is to answer a support ticket and close it. If it gets stuck or isn't sure about an answer, it's supposed to escalate the ticket to a human.

  • Knowledge Base Agent: This agent finds common questions in your support tickets that don't have a corresponding help article and drafts one for you to review.

  • Prospecting Agent: It can do things like enroll contacts into an email sequence and send the follow-ups for you.

Automated email follow-up in HubSpot Service Hub, showcasing Breeze Agent inputs and outputs in action.::
Automated email follow-up in HubSpot Service Hub, showcasing Breeze Agent inputs and outputs in action.

The key thing to remember is that these actions are stuck inside HubSpot. If you need the agent to do something in an external system, like checking an order status in your Shopify store, you're out of luck without custom development. This is a key difference from more flexible platforms. With a tool like eesel AI, for example, you can set up custom actions that let the AI pull information or do things in any of your other business tools. That opens up a lot more possibilities.

The catch: Key limitations

While the tight integration with HubSpot has its perks, it also comes with some serious limitations you should know about before committing.

  1. You're locked into one ecosystem: If your business doesn't run 100% on HubSpot, the agents will have major blind spots. They can't see knowledge in other tools, and they can't take action outside of HubSpot. This makes them a tough fit for companies that use a variety of best-in-class tools for different jobs.

  2. Not much room for testing or fine-tuning: HubSpot designed its agents to be simple, but that simplicity means you give up a lot of control. It's hard to be specific about which tickets an agent should work on or to really test its performance before you let it loose on your customers. There's no "simulation" feature to see how it might have handled last month's tickets.

  3. A rigid, all-or-nothing approach: Because the agents are a built-in feature, you're on HubSpot's schedule for updates and new capabilities. You can't just start with a small, simple use case and build from there. It's a big step to take all at once.

For teams that want to be sure before they go live, eesel AI has a really neat simulation mode. It lets you run the AI on your past support tickets to see exactly how it would have performed. You get a clear, data-backed forecast of its resolution rate and can spot any knowledge gaps before it ever talks to a real customer. You can then roll it out slowly, maybe starting with just one type of ticket, and expand as you get more comfortable.

The price tag: How much do Breeze agents cost?

So, how much does this all cost? You can't just buy Breeze Agents as a cheap add-on. They're only included in HubSpot's more expensive plans, so you won't find them on the Free or Starter tiers.

Agent TypeRequired HubPlan LevelMonthly Price (Annual Contract)
Content & Social MediaContent Hub or Marketing HubProfessional$800/mo
Enterprise$3,600/mo
Prospecting AgentSales HubProfessional$1,200/mo
Enterprise$5,000/mo
Customer & KB AgentsService HubProfessional$1,200/mo
Enterprise$5,000/mo
  • Content & Social Media Agents: You'll need Content Hub or Marketing Hub at the Professional ($800/mo) or Enterprise ($3,600/mo) level.

  • Prospecting Agent: This requires Sales Hub at the Professional ($1,200/mo) or Enterprise ($5,000/mo) level.

  • Customer & Knowledge Base Agents: You need Service Hub at the Professional ($1,200/mo) or Enterprise ($5,000/mo) level.

Keep in mind these prices are for annual contracts and can go up if you have more contacts or users. On top of that, the Breeze Intelligence feature that enriches your data runs on a separate credit system, adding another cost to the pile. This makes Breeze Agents a pretty hefty investment, best suited for companies that are already all-in on the HubSpot platform.

Are Breeze agents the right call for your team?

At the end of the day, HubSpot's Breeze Agents are a solid example of what happens when you bake AI directly into a CRM. If your company is already heavily invested in the HubSpot Professional or Enterprise plans, they offer a convenient way to automate work without needing another tool. Having direct access to your CRM data makes for some pretty smart content and actions that a generic AI tool couldn't pull off.

But the thing that makes them convenient is also their biggest drawback. Being locked into the HubSpot ecosystem means you lose a ton of flexibility. You're limited in how you can control them and can't easily connect them to knowledge stored in your other tools. The high price tag and lack of a good way to test them beforehand make it a big, and slightly risky, commitment.

If your team needs something more flexible, a tool that gives you more control and lets you test things out without risk, you'll probably be better off with a solution built to work with your entire tech stack. A platform like eesel AI is self-serve, so you can get it running in minutes. It connects to all your knowledge sources right away and lets you test its performance before you flip the switch. It works with the help desk you already have, whether that's Zendesk, Freshdesk, or something else, so you don't have to change your whole setup.

Ready to see what a truly flexible AI agent can do? Try eesel AI for free and build your first AI support agent in minutes.

Frequently asked questions

The inputs for Breeze Agents primarily consist of your HubSpot CRM data, knowledge base content, and past marketing materials. The outputs are generated content (like blog posts or social media updates) and automated actions within HubSpot, such as answering support tickets or enrolling contacts in sequences.

Effective Breeze Agent Inputs and Outputs heavily rely on robust CRM data (contacts, companies, deals), a comprehensive HubSpot knowledge base, and your historical content (blogs, landing pages) to learn brand voice and context. Without clean and complete data, the agent's performance will be limited.

A significant limitation of Breeze Agent Inputs and Outputs is their confinement to the HubSpot ecosystem. They cannot access external knowledge bases or take actions in non-HubSpot tools, which creates blind spots if your data is dispersed across various platforms.

To utilize the Breeze Agent Inputs and Outputs, you generally need HubSpot's Professional or Enterprise plans for specific hubs. For instance, Content and Social Media Agents require Content Hub or Marketing Hub Professional/Enterprise, while customer agents need Service Hub Professional or Enterprise.

For customer support, the Breeze Agent Inputs and Outputs allow the Customer Agent to process incoming tickets using your knowledge base. It can answer common questions and close tickets, escalating to a human only when it's unable to resolve an issue or uncertainty arises.

Teams can accomplish tasks such as drafting blog posts, creating social media updates, personalizing sales emails, answering customer support tickets, and drafting new knowledge base articles using the Breeze Agent Inputs and Outputs. These actions are all managed within the HubSpot platform.

The blog mentions that HubSpot's agents offer limited fine-tuning and no direct "simulation" feature to test Breeze Agent Inputs and Outputs performance before live deployment. This means it's a more rigid, all-or-nothing approach to rolling them out into your live environment.

Share this post

Kenneth undefined

Article by

Kenneth Pangan

Writer and marketer for over ten years, Kenneth Pangan splits his time between history, politics, and art with plenty of interruptions from his dogs demanding attention.