What This Update Actually Is
The Breeze Knowledge Base Agent is a new AI agent inside HubSpot's Content section. It connects to your ticket pipeline, reads closed tickets, and compares what customers actually asked against what your knowledge base already covers.
When it finds a gap, it doesn't just flag it. It writes a draft article and queues it for your review. You read it, edit if needed, and publish through the normal article editor.
This is currently in private beta. It's available to Service Hub Pro and Service Hub Enterprise portals. You won't see it if you're on Starter or Free.
Why HubSpot Shipped This
Here's the real problem. Most support teams keep a knowledge base, but they never quite keep up with it. Tickets pile up. Reps answer the same questions over and over. And the documentation sits months behind the actual product or process.
The internal frustration is real: someone on your team knows exactly what should be documented, but writing it feels like the task that never makes it to the top of the queue. Meanwhile, your customers submit another ticket asking the same thing they asked last month.
HubSpot is betting that if they remove the blank-page problem, teams will actually maintain their knowledge bases. The agent handles the drafting. Humans handle the judgment call on accuracy and tone.
How to Use It Step by Step
Before you start, confirm you meet the requirements:
- Your portal is on Service Hub Pro or Enterprise
- You have at least 5 closed tickets in the pipeline you want to pull from
- You're on HubSpot's newest knowledge base tools with a knowledge base already set up
Once you've confirmed those, here's how to get started:
- Navigate to Content, then select Knowledge Base in the left menu.
- Click the Knowledge Base Agent tab. If you see it, you're in the beta.
- Click "Create knowledge base agent" and select your knowledge base and your ticket pipeline.
- Click Finish. The agent will begin analyzing ticket data against your existing articles.
- When drafts appear, click "View sources" to see which tickets and gaps informed each article.
- Click any article title to open the draft in the article editor. Edit as needed, then publish normally.
- Use the feedback option in the editor to rate the AI-generated content. This helps the agent improve over time.
One thing to note: the agent produces articles in the primary language set on your knowledge base. If you serve multiple languages, make sure the correct language is set before you configure the agent.
What It Touches in Your HubSpot Strategy
This update lives inside Service Hub, but it ripples outward in a few meaningful directions.
First, your ticket data becomes a content asset. Every closed ticket is now potential documentation. The quality of your ticket data, meaning clear subject lines, resolved status, and accurate pipelines, directly affects the quality of the drafts the agent produces.
Key Takeaway
Messy ticket data produces messy drafts. Before you run the agent, make sure closed tickets have descriptive subjects and are sitting in the right pipeline. Garbage in, garbage out still applies to AI.
Second, this connects your support workflow directly to your self-service content strategy. If your team is already using HubSpot's Customer Agent or chatbot tools to deflect tickets, a stronger knowledge base means better deflection rates. These two systems feed each other.
If you've been working through HubSpot's recent service changes, including the legacy channel routing sunset in Customer Agent, this update fits into the same broader push: get humans out of repetitive service loops and let AI handle the predictable volume.
Third, this is a content operations shift. Whoever owns your knowledge base, whether that's a support manager, a content strategist, or an operations lead, now has a review-and-approve workflow instead of a write-from-scratch workflow. That's a different skill set and a different time investment.
Key Takeaway
The agent doesn't replace editorial judgment. Someone still needs to verify that each published article is accurate, on-brand, and safe to surface to customers. Build a review step into your publishing process before you go live.
If your team is thinking about the full arc of how customers experience your brand after the sale, this update belongs in that conversation. A well-maintained knowledge base is one of the highest-leverage touchpoints in the post-purchase journey. We go deeper on that in our piece on why most B2B customer journey maps fail.
Who Should Care Most
This update matters most to specific roles and company profiles. Here's who should move on it first:
- Support managers at Service Hub Pro or Enterprise portals who have a knowledge base that hasn't been updated in months. The agent gives them a shortcut to catching up.
- Content strategists who own knowledge base governance but don't have bandwidth to write every article from scratch. Review workflows are faster than blank-page workflows.
- RevOps leaders who want to reduce ticket volume without hiring more support reps. Better self-service documentation is a cost-reduction lever.
- Growing companies whose support team is still small. If two humans are handling all tickets, this agent scales the documentation side without adding headcount.
This update won't move the needle for teams with fewer than 5 closed tickets or teams that don't yet have a knowledge base set up. Get those foundations in place first.
George's Take
“The blank page has always been the enemy of good documentation. HubSpot just removed it from the equation for support teams. But I want to be honest: the agent is only as good as your ticket hygiene and your reviewer's judgment. Don't publish every draft without reading it. Use this to get to 80%, then bring your humans across the finish line.”
What I find genuinely exciting here isn't just the time savings. It's that your support conversations finally become a knowledge asset that compounds over time. Every ticket your team closes is now potentially improving the experience for the next customer who hits the same wall. That's a flywheel most companies never build because writing feels too heavy. The Breeze Knowledge Base Agent makes it lighter, but your team still has to decide what's worth publishing. That judgment is yours. Keep it.
This fits a larger pattern we're tracking across Breeze. If you haven't caught up on how Breeze Assistant now works inside Slack, that's another piece of the same story: AI reducing the friction between your data and the humans who need to act on it.
If you want help setting up your knowledge base the right way before this agent goes live in your portal, or if you're not sure your ticket pipeline is clean enough to produce useful drafts, let's talk. Book a strategy call with the Sidekick team at sidekickstrategies.com and we'll walk through your portal together.
Frequently Asked Questions
What is HubSpot's Breeze Knowledge Base Agent?
The Breeze Knowledge Base Agent is an AI tool inside HubSpot's Service Hub that reads your closed support tickets, identifies topics your knowledge base doesn't cover, and drafts new articles automatically. You review, edit, and publish the drafts using the standard HubSpot article editor. It's currently in private beta for Service Hub Pro and Enterprise.
What do I need before I can use the HubSpot Knowledge Base Agent?
You need a Service Hub Pro or Enterprise subscription, at least 5 closed support tickets in the pipeline you want the agent to analyze, and an existing knowledge base set up on HubSpot's newest knowledge base tools. Without these three things in place, the agent tab won't be available in your portal.
Does the Breeze Knowledge Base Agent publish articles automatically?
No. The agent drafts articles for your review. You must click into each draft, read it, make any edits, and publish it manually through the article editor. HubSpot also lets you view the source tickets that informed each draft so you can verify the content before it goes live.
Which language will the Knowledge Base Agent write articles in?
The agent produces articles in the primary language set on the knowledge base you select during setup. If you manage a multilingual knowledge base, confirm the correct primary language is configured before you create the agent. You may need to run separate agents for separate language-specific knowledge bases.
How does the Knowledge Base Agent know what topics are missing?
The agent compares the topics and questions found in your closed support tickets against the articles already published in your knowledge base. Where it finds gaps, meaning questions that tickets resolved but your docs don't address, it treats those as candidates for new articles and generates a draft for each one.
Will this reduce my support ticket volume?
It can, over time. A more complete knowledge base gives customers the ability to self-serve before submitting a ticket. Teams that consistently maintain their knowledge base with answers to common questions typically see deflection rates improve. This agent makes it easier to keep your knowledge base current, which is the underlying requirement for deflection to work.
