HubSpot Breeze Content Agent: A Practical Guide

Most marketing teams running HubSpot now have an AI content engine sitting inside their portal, whether they have switched it on or not. The Breeze Content Agent is HubSpot's autonomous helper for producing marketing content at scale, and it has matured quickly from a single button into a set of focused tools for blog research, case studies and more. If your team is still hand-writing every meta description and starting each blog from a blank page, you are leaving a good deal of value on the table.

This guide explains what the Content Agent actually does, where it genuinely shines, how to get real value from it rather than generic output, and where a custom build can take you further. It sits within our wider view of AI for HubSpot, and pairs well with the Breeze agents overview if you want the bigger picture across HubSpot's agent family.

What the Breeze Content Agent is

The Content Agent is one of HubSpot's autonomous Breeze Agents. It is distinct from the Breeze Assistant (the in-app conversational AI, sometimes called Copilot, that responds to you in the moment grounded in your CRM) and from Breeze Intelligence (data enrichment and buyer intent). The Assistant answers questions and helps you as you work. The Content Agent does something different: it scales content production, generating drafts from your business context, brand style and target audience.

HubSpot positions the agent to create content across blogs, landing pages, case studies and podcasts. In its enhanced form launched in Spring 2025, HubSpot states the agent can use uploaded reference files for blog posts and landing pages, suggest stronger blog topics based on your top-performing content and audience, and automate pre-publish tasks such as writing meta descriptions, creating confirmation emails on form submissions, and adding internal links.

One thing to understand up front is that the packaging is in flux. The functionality increasingly appears in HubSpot's product as discrete tools rather than one monolithic "Content Agent" button. That includes a Blog research agent (which researches topics, identifies sources and generates optimised long-form posts) and case study generation, found via the "Start with AI" option under Content and then Case Studies. HubSpot's Content Hub product page markets these collectively as built-in AI agents that turn your strategy into multi-channel content. Expect feature names and availability to keep shifting, so treat any single screenshot as a snapshot rather than a fixed contract.

Case study generation is a good example of how structured the workflow has become. It takes a topic or idea, a target-audience country, an optional company record and ICP, a content language, and uploaded supporting documents (audio, image, text or video). It then proposes SEO-scored title options, complete with monthly search volume, ranking difficulty and keyword intent, before drafting the full case study. Podcast support, per HubSpot's Content Hub marketing, extends to scripts, episode outlines, show notes and artwork, and can repurpose existing assets such as webinar recordings or blog content into episodes. Audio and voice specifics are evolving, so verify the current state before you build a process around them.

Where it shines

For teams already invested in HubSpot, the Content Agent is a sensible and often excellent first move. Because it lives inside Content Hub, it draws on your CRM context, your existing content and your brand materials without any integration work. There is no separate add-on to buy and no data pipeline to stand up. That tight coupling is its biggest practical advantage.

It is genuinely strong at the unglamorous, repetitive work that quietly slows teams down. Writing meta descriptions, drafting form confirmation emails and adding internal links are exactly the tasks that get skipped under deadline pressure, and they matter for both SEO and user experience. Handing them to the agent frees your writers to do the work that needs a human.

The Blog research agent is also a credible answer to the blank-page problem. By grounding topic suggestions in your top-performing content and your target audience, it points you towards subjects you have earned the right to write about, rather than generic ideas pulled from nowhere. And because reference files anchor the output, drafts can sound like your brand rather than reading as off-the-shelf AI. Crucially, a human editor stays in charge by design. Drafts and suggestions land in the standard editor and drafts workflow. Blog research agent suggestions appear in a "Latest Agent Suggestions" area and only move to drafts when opened in the editor. Nothing publishes without a human action, which is the right default for anyone who cares about quality and brand safety.

Getting real value from it

The single biggest reason teams are disappointed by the Content Agent is that they switch it on and expect magic. Output quality is only as good as the inputs. If your reference files are thin or out of date, drafts read generic, and people conclude the tool does not work. The fix is preparation, not more prompting.

Start with access and setup. The Content Agent and Blog research agent require the right edition: Content Hub Professional or Enterprise (the Blog research agent is also available on Smart CRM Professional or Enterprise). A Super Admin needs to enable Breeze and the relevant AI data settings, such as CRM data and customer conversion data, and an existing blog must already be created for the Blog research agent to work against. Users also need edit and publish permissions. None of this is a separate add-on, but it does need configuring deliberately.

Then invest in your knowledge base. Audit your brand and tone guidelines, gather your best-performing past blogs, and pull together the technical PDFs and product documents that hold your real expertise. Upload these as reference files so the agent has strong material to draw on. This is the step most teams underuse, and it is the one that separates on-brand drafts from forgettable ones.

Set guardrails too. Super Admins can configure extra instructions, access permissions and automation triggers in Breeze Studio. In 2026, Breeze Agents can also be triggered from a HubSpot workflow via the Run Agent action, currently in beta, which lets you fold agent steps into an existing process. Keep humans firmly in the loop: let the agent handle topic research, first drafts and pre-publish housekeeping, then have an editor add the experience, judgement and first-hand E-E-A-T signals that AI cannot supply.

A few limits are worth planning around. Breeze content generation carries documented usage limits: HubSpot's Knowledge Base notes 30 generations per minute, 1,000 per day and a 4,000-token cap per prompt and response. Agent actions can also consume HubSpot credits. None of this is a blocker for normal use, but it is worth knowing before you script a high-volume process.

Where a custom Claude agent goes further

The Content Agent is excellent within its boundaries, and being honest about those boundaries is the point. It is bounded to HubSpot's CRM and Content Hub context, to the editions you hold, to credit and usage limits, and to templated workflows. For a large share of marketing content, that is exactly what you want.

Some work sits outside those edges. If your content needs to draw on a private knowledge base or reference corpus that does not live in HubSpot, if you need to enforce bespoke editorial rules and multi-step review chains, if you want to choose the model or integrate non-HubSpot tools and data, or if you need long-context, deeply tailored output, a custom build is the better fit. A custom agent built on Anthropic's Claude, orchestrated by a consultancy, can ingest arbitrary private sources, apply your own governance and produce automation that simply sits outside HubSpot's packaging.

This is a complement, not a competition. HubSpot's own Run Agent workflow action means a Breeze agent and a custom Claude agent can run side by side in the same process. The realistic picture for most growing teams is Breeze doing the high-volume, in-platform content work, with a custom agent handling the specialist, cross-system or tightly governed jobs Breeze was never designed for.

How SpotDev helps

SpotDev is a HubSpot Diamond Partner and a Claude specialist, and we do both deliberately. We implement, configure and optimise the Breeze Assistant and Breeze Agents so you get real value from HubSpot's own AI, including the setup, reference files and guardrails that most teams skip. We also have in-house engineers who design and build custom Claude agents for the work Breeze cannot do, wired into HubSpot. Nothing is subcontracted, and we have delivered 300+ technology projects.

Because we do both, we can be an honest broker. We will tell you plainly when Breeze is enough and when a custom build pays back, rather than steering you towards whichever one we happen to sell. If you want to weigh up your options, you can talk to a Claude-specialist engineer who knows both sides of the picture.

Frequently asked questions

Do I need a specific HubSpot edition to use the Breeze Content Agent?

Yes. The Content Agent and Blog research agent require Content Hub Professional or Enterprise, and the Blog research agent is also available on Smart CRM Professional or Enterprise. It is not a separate add-on. A Super Admin must enable Breeze and the relevant AI data settings, and for the Blog research agent you need an existing blog already created in HubSpot.

Does the Content Agent publish content on its own?

No. It produces drafts and suggestions that land in HubSpot's standard editor and drafts workflow. Blog research agent suggestions appear in a "Latest Agent Suggestions" area and only move to drafts when you open them in the editor. A human has to review, refine and approve before anything goes live, which keeps an editor in control of quality and brand.

Why does my AI-generated content read as generic?

Almost always because the inputs are thin. Output quality depends heavily on your reference files. If you upload strong brand and tone guidelines, your best-performing past content and the technical documents that hold your real expertise, drafts sound like you. Without that material, the agent has little to work from and the writing reads off-the-shelf.

When should we use a custom Claude agent instead of Breeze?

Use Breeze for high-volume content work that lives inside HubSpot. Consider a custom Claude agent when the work spans systems beyond HubSpot, needs bespoke editorial logic or stricter governance, requires a specific model or long-context handling, or must draw on private knowledge bases. The two can run side by side, including via HubSpot's Run Agent workflow action.

Work with a Claude specialist

SpotDev designs, builds and deploys custom Claude agents and enterprise Claude rollouts for UK businesses, with fixed packages from £8,000 to £45,000 and a first rollout live in two to three weeks. Explore our Claude implementation packages or talk to one of our engineers.