What AI does HubSpot use? A complete guide to Breeze AI

It seems like every software tool slapped an "AI" badge on its features over the last year. It’s changing how we work, especially when it comes to managing customer relationships. HubSpot, being one of the biggest names in the CRM world, jumped right in with its own AI suite, now called "Breeze." It promises to make everything from marketing to sales and service a whole lot smarter.

But with all the hype, it’s easy to feel a bit lost. What is HubSpot’s AI, really? What’s going on under the hood? And more importantly, what can’t it do? Is it actually the right fit for your team, or are there better options out there?

This is your no-fluff guide to HubSpot’s Breeze AI. We’ll break down what it is, the AI models it runs on, its main features, and where it just doesn’t quite measure up, especially when you compare it to more focused, flexible AI tools.

What AI does HubSpot use, exactly?

Alright, let’s get right to it. When you ask, "what AI does HubSpot use?", the answer isn’t a single thing. HubSpot takes a mix-and-match approach, blending its own homegrown AI with some of the heaviest hitters from other companies.

According to HubSpot’s own AI Trust Center, they’re using a handful of different models to power Breeze. This includes tech from:

• OpenAI: They’re using various models, including the latest GPT-4o. These are the brains behind complex tasks and generating text that sounds like a real person wrote it.

• Anthropic: The Claude family of models is in the mix too, known for being great at conversational AI and creating content.

• Stability AI: For visuals, like whipping up a quick blog header or a social media graphic, they tap into models from Stability AI.

This multi-model strategy is pretty smart. It lets HubSpot pick the right tool for the job, whether that’s drafting a sales email or digging through your CRM data for insights. For you, all this tech is just neatly bundled together under the Breeze AI brand.

Understanding Breeze AI: HubSpot’s AI suite

Breeze AI is basically the brand name for every AI-powered feature you’ll find across the HubSpot platform. The whole idea is to weave AI into the daily grind for marketing, sales, and service teams. It’s not one single tool, but a collection of features that fall into three main categories.

Breeze Copilot: The AI sidekick

Think of the Breeze Copilot as your helpful assistant that follows you around inside HubSpot. It’s a conversational tool designed to help you with all the little things. You can ask it to summarize a long email chain with a customer, draft a quick follow-up, or pull key details from a contact record so you’re prepped for a call.

It pops up across the Marketing, Sales, and Service Hubs, offering help that’s relevant to what you’re doing at that moment. Its main purpose is to help you spend less time on tedious tasks and more time on work that actually requires your brain.

Breeze Agents: Automating workflows

Breeze Agents are a step up from the Copilot. They aren’t just assistants; they’re set up to be more like autonomous "teammates" that can handle workflows with multiple steps. For instance, the Content Agent can generate blog posts, the Prospecting Agent can dig up info on leads and write personalized outreach, and the Customer Agent can handle initial support questions.

These sound great, but they come with one big catch: they are designed to work entirely within the HubSpot universe. This becomes a major headache if your team’s most important knowledge isn’t in HubSpot. What if your wikis are in Confluence, your support playbook is in Zendesk, and your team talks through problems in Slack? The agents can’t see any of it.

Breeze Intelligence: Enriching CRM data

Breeze Intelligence is all about making the data you already have even better. It automatically adds information to your contact and company records from other sources, a feature that got a serious boost when HubSpot bought the data company Clearbit.

Its main tricks are figuring out which companies visiting your website are actually ready to buy and automatically shortening your lead forms to get more people to fill them out. The goal is to make your CRM a living, breathing source of truth that helps you make smarter moves.

A closer look at HubSpot’s AI tools in action

It’s one thing to hear about copilots and agents, but what does it actually feel like to use HubSpot’s AI? Let’s look at how it works for different teams.

For marketing and content creation

HubSpot’s AI is pretty handy for content work. Tools like the AI Blog Writer and AI Email Writer are perfect for getting started. They can help you brainstorm outlines, spit out a first draft, and push past that blinking cursor on a blank page.

That said, some users find the tools can be a bit stubborn. You’ll see discussions in online forums about how the AI sometimes struggles with specific instructions. For example, you might give it perfectly good copy for a landing page, but instead of using it or building on it, the AI just overwrites it with its own version. That lack of control can be a real pain when you know exactly what you want.

For sales and lead management

For sales folks, HubSpot’s AI is all about working faster. The predictive lead scoring feature looks at your data to flag which leads are most likely to turn into customers, helping your reps focus their energy in the right place. You can also use the AI reporting assistant to build dashboards and pull out insights without having to spend hours digging through spreadsheets.

For customer service

On the support side of things, HubSpot’s AI Chatbot and Customer Agent are the key features. They learn from your HubSpot knowledge base articles and your public website to answer common customer questions.

This sounds fine on the surface, but it misses the most valuable source of knowledge your support team actually has: the thousands of successfully resolved tickets sitting in your helpdesk. This is where a specialized tool like eesel AI makes a world of difference. It connects directly to the helpdesks you already use, like Zendesk or Freshdesk, and learns from your team’s past conversations. This means its answers aren’t just pulled from generic help articles; they reflect how your team actually solves tricky customer problems, making the support way more accurate and genuinely helpful.

The eesel AI Copilot drafting a personalized response within a helpdesk, showcasing how it learns from past tickets to provide accurate support.

The hidden costs and limitations

The convenience of having AI built right into your main platform is hard to ignore, but tools like HubSpot’s come with some pretty big trade-offs. It’s worth understanding these before you go all in.

The "walled garden" problem

HubSpot’s AI is built to work with HubSpot’s data. That’s both its greatest strength and its biggest weakness. It creates a "walled garden" where the AI is powerful, but only as long as every last piece of your company’s information lives inside HubSpot. But let’s be realistic, whose does? What if your internal documentation is a bunch of Google Docs? Or your team’s collective brain lives in Confluence or Slack? HubSpot’s AI is blind to all of it, leaving it with huge knowledge gaps.

A platform like eesel AI, on the other hand, is built specifically to connect all your scattered knowledge. With one-click integrations for over 100 tools, it pulls everything together without making you go through a massive, painful data migration project just to get your AI up and running.

An infographic comparing HubSpot's "walled garden" AI with eesel AI's ability to integrate knowledge from over 100 external sources.

Lack of deep customization and control

As we touched on earlier, HubSpot’s AI tools can feel a bit rigid. You get a set of features, but you have to use them the "HubSpot way." This is tough if your team has unique workflows or you need an AI that can handle very specific tasks.

This is where a tool that puts you in the driver’s seat really shines. For example, eesel AI’s agent builder gives you complete control. Its visual workflow editor lets you decide exactly which tickets get automated and which ones need a human touch. You can tweak the AI’s personality with a simple prompt and even connect it to external APIs to do things like look up order details from your Shopify store on the fly.

The eesel AI agent builder interface, showing the visual workflow editor that allows for deep customization and control.

No way to test the AI before you go live

Letting an AI start talking to your customers is a big step. You want to be damn sure it’s going to do a good job and not make your company look bad. With HubSpot, there isn’t really a way to test how the AI will perform before you unleash it on your customers. You’re more or less flipping a switch and hoping for the best.

This is a huge advantage of a dedicated support automation tool. eesel AI comes with a powerful simulation mode that lets you safely test your AI setup on thousands of your past support tickets. You can see exactly how it would have answered, get a solid prediction of its resolution rate, and spot any gaps in its knowledge. It lets you launch feeling confident, not anxious.

The eesel AI simulation dashboard, which predicts automation rates and shows how the AI would have handled past customer tickets.

Pricing and what to expect

HubSpot’s pricing for its AI features is pretty typical for HubSpot: you get some basic tools on the free and starter plans, which is a nice way to dip your toes in.

But the really powerful stuff, like the Breeze Agents, is kept for the pricey premium tiers (Professional or Enterprise plans), which come with some serious per-seat costs. On top of that, Breeze Intelligence is often sold as a separate add-on that works on a credit system. This can make your monthly bill a bit of a guessing game, with costs jumping up when you’re busy.

It’s a big difference from the straightforward pricing of a tool like eesel AI. Their plans are based on a predictable monthly volume, and you don’t have to worry about surprise fees for every ticket it solves. What you see is what you pay.

A screenshot of eesel AI's clear, volume-based pricing page, contrasting with HubSpot's more complex, multi-tiered pricing model.

The verdict: Is HubSpot’s AI right for you?

So, after all that, should you use HubSpot’s AI? The answer really depends on your company’s setup.

HubSpot’s Breeze AI is a decent, convenient option if your company is already living and breathing HubSpot. If your teams work almost exclusively inside that ecosystem and you just need some general AI tools to make internal tasks a bit faster, it’s a solid choice.

However, if your main goal is to build a smart, accurate, and super-customizable AI for your customer support that works with all the tools you already use, then a specialized platform is going to be a much better bet. You need something that can learn from your helpdesk history, connect to your wikis, and give you full control over how it behaves.

If you want to automate your support, bring all your knowledge sources together, and have total command over your AI without being locked into one platform, you should give eesel AI a try.