What This Update Actually Is
HubSpot shipped automatic topic classification for Help Desk and Customer Agent. Once topic tracking is on, HubSpot classifies eligible tickets and conversations by topic. No rules to write. No training data to prep. No rep behavior to police.
You start with a default taxonomy out of the box. You can extend it with custom topics and subtopics whenever your business needs more precision. Reps can also manually override a ticket's primary topic or its full topic list if the auto-classification misses the mark.
Two new surfaces show you the data. In Help Desk, a Topics view shows each topic's number of tickets, CX Score, Overall Communication Score (QA), and Time to Close SLA Adherence. In Customer Agent, the Topics view shows where the AI agent is struggling to answer certain kinds of questions well. Both are available to Service Hub Pro and above.
New ticket properties for topics and topic drilldowns are also available in custom reports, so this data travels into whatever reporting stack you've already built.
Why HubSpot Shipped This
Support leaders face a problem that sounds simple but rarely gets solved cleanly: what are customers actually reaching out about, and where is the team falling short?
The old answer was manual ticket categories. That answer has two failure modes. First, reps tag inconsistently, especially during high-volume periods. Second, category lists go stale fast. A new product line launches, a bug ships, a billing change confuses humans, and the category list doesn't catch up for weeks.
The result is a reporting picture that's always slightly wrong. You make coaching decisions on bad signal. You miss emerging issues until they become escalations. You run a data project every quarter just to get a baseline read.
HubSpot's automatic topic classification removes the dependency on rep behavior entirely. The classification happens at the conversation level, not the ticket level alone, which means even handoff conversations and Customer Agent sessions contribute to the picture.
How to Use It Step by Step
- Open Help Desk and navigate to the Topics view. You'll see your ticket volume organized by topic automatically. First, a Super Admin opts into the Topics beta and turns on topic tracking. It may take some time for topics to appear.
- Compare topics by number of tickets, CX Score, Overall Communication Score (QA), and Time to Close SLA Adherence to find your highest-friction topics. These are the areas where customers are struggling most, and where your reps are spending the most time.
- Open Customer Agent's Topics view to see where the AI agent is underperforming by topic. If the agent is consistently failing on billing questions, for example, that's a knowledge gap you can fix in the agent's configuration.
- Review the default taxonomy. Add custom topics and subtopics for anything the default set doesn't cover. Keep custom topics specific enough to be actionable but broad enough to accumulate enough volume to be meaningful.
- Build custom reports using the new topic ticket properties. Pull topic data into existing dashboards so leadership sees the same picture the support team does.
- Use the manual override when a rep spots a misclassification. This keeps your data clean without requiring a full re-training workflow.
What It Touches in Your HubSpot Strategy
This update is more than a reporting feature. It changes how several connected pieces of your HubSpot setup should work.
On the CRM properties side, topic properties now live on ticket records. That means you can segment tickets by topic in list views, trigger workflows based on topic, and use topic as a filter in any report. If your ops team has built ticket-based automation, topics give you a new and much more reliable trigger condition than manual category fields.
On the Customer Agent side, this pairs directly with the agent's handoff behavior. If the agent is struggling on a specific topic, you now have the data to see it clearly before it becomes a volume problem. Think of topic analysis as the feedback loop that makes your agent smarter over time.
If you're using the feature where Customer Agent collects ticket properties before handing off to a human, topic data now adds another dimension to what gets captured at handoff. You can see not just what properties were collected, but what kind of issue triggered the handoff.
Key Takeaway
Topic properties on ticket records mean you can now trigger workflows, build segments, and filter reports by what customers are actually asking about, not by what a rep remembered to tag.
On the reporting side, this is new signal you didn't have before. You can now tie ticket topic to CSAT scores, resolution time, rep performance, and even revenue impact if you've connected service data to deals. That connection matters a lot for RevOps teams trying to show how support quality affects retention.
If your team is already exploring how service data connects to revenue, the Revenue Reporting from Contracts update is worth reading alongside this one. Topic-level support performance and contract-level revenue data are the two sides of the retention story.
Key Takeaway
Coaching decisions made on topic-level data are more durable than coaching made on gut feel or one-off ticket reviews. This update gives managers a repeatable, evidence-based coaching rhythm.
This feature also fits into the broader direction HubSpot is moving with its agentic platform. If you want the full picture of where AI-powered service is headed, our HubSpot Agentic Platform deep dive gives you the strategic context.
Who Should Care Most
This update matters most to four groups of humans.
- Support managers who want to coach on data, not instinct. Topics give you a consistent weekly signal for where to focus 1:1s and team training.
- RevOps leaders who need reliable ticket categorization for forecasting, capacity planning, and QBR reporting. Manual tags are notoriously unreliable as a data source. Auto-classification fixes that at the source.
- Product and CX teams who want to catch emerging issues fast. When a new bug or confusing workflow triggers a spike in a specific topic, you'll see it in volume trends before it hits escalations.
- Owners of growing companies on Service Hub Pro who want their support function to run with less management overhead. The Topics view does the analysis work so you don't have to run a report to see where the pain is.
HubSpot's help article lists topics for Service Hub Professional and Enterprise only, so don't count on topic classification if your portal is on Service Hub Starter or Free.
George's Take
I've audited a lot of HubSpot portals, and the support data situation is almost always the same story. The category field exists. Nobody trusts it. Leadership stops pulling reports from it. Then coaching becomes reactive, capacity planning is a guess, and nobody can tell you with confidence what the top five reasons humans contact support actually are. That's a fixable problem, and HubSpot just fixed the root cause instead of patching around it. Automatic classification means the data is as good as the AI's read on the conversation, not as good as the busiest rep's memory at ticket close. That's a meaningful shift.
“When you stop depending on human consistency for data quality, the data actually becomes useful. That's what this update does for support teams.”
If your support team is running on Service Hub Pro or Enterprise and you're not already using Customer Agent and a structured Help Desk setup, now is a good time to get that foundation right. The Topics feature compounds in value the more traffic flows through it. A portal that's only half-using Help Desk gets half the picture.
If you're also trying to reduce the time managers spend scanning notifications and activity, the On-Demand AI Summary for the HubSpot Notification Sidebar pairs well with this update. Less noise, clearer signal, faster decisions.
Want to make sure your Help Desk setup is actually capturing the full value of features like this? Book a strategy call with the Sidekick team. We'll walk through your current portal setup and show you exactly where you're leaving performance data on the table.
Frequently Asked Questions
Does HubSpot's topic classification require any manual setup or training data?
There's no training data to prep, but there is one setup step: a Super Admin opts into the Topics beta and turns on topic tracking. It may take some time for topics to appear. You start with a default taxonomy out of the box. You can add custom topics and subtopics at any time, and reps can manually override a classification if the AI gets it wrong. No rules engine, no training pipeline required.
Which HubSpot tier do you need to access the Topics view in Help Desk and Customer Agent?
Topics require Service Hub Professional or Enterprise, and you need a Service Hub Professional or Enterprise seat to access topics in Help Desk. HubSpot's help article doesn't list Free or Starter.
Can topic data be used in HubSpot workflows or custom reports?
Yes. HubSpot added new ticket properties for topics and topic drilldowns. These properties are available as workflow enrollment triggers and conditions, and they're fully accessible in the custom report builder. That means you can automate escalations by topic or build executive dashboards that track support performance by issue type.
How does topic analysis help with coaching Customer Agent?
Customer Agent's Topics view shows which topics the AI agent handles poorly, measured by resolution quality and escalation patterns. If the agent consistently struggles with billing or onboarding questions, you can identify that gap specifically and update the agent's knowledge base or instructions to address it.
What happens if a ticket covers more than one topic?
HubSpot assigns a primary topic and also tracks all topics mentioned in a conversation. Reps can manually update both the primary topic and the full list of associated topics. This means multi-issue tickets are handled accurately and your reporting doesn't artificially collapse complex conversations into a single category.
Is this the same as HubSpot's existing ticket category field?
No. The existing category field is a manual property that reps fill in. Topic classification is automatic, AI-driven, and applied at the conversation level. The new topic properties are separate from any custom category fields you've built. You can run both in parallel while you validate the classification quality before retiring manual tagging.
