Find out what your tickets are really telling you

What: This prompt asks Breeze Assistant (HubSpot's built-in AI assistant) to read your recent tickets - the records HubSpot creates for every customer support request - and group them by root cause. For each theme, it names the team best placed to stop those tickets arriving and suggests one specific fix. It replaces the usual monthly service report, which tells you how fast you answered, with a short list of problems other teams can actually solve.

Your service metrics measure coping. Time to first response, tickets closed, and backlog size tell you how well the team handles demand. None of them explain why the demand exists. You can hit every SLA (service level agreement) for a year and still receive the same 60 tickets a month about the same invoice.

The fix usually lives in another team. In most portals we audit, the biggest ticket themes trace back somewhere else: a quote template that leaves out VAT, an onboarding plan that skips user permissions, or a sales deck promising an integration that doesn't exist yet. Support sees the fallout, but it can't change any of those things on its own.

HubSpot will summarise your tickets, but it won't route them. Breeze Assistant can summarise individual ticket records, and HubSpot's Knowledge Base agent can turn resolved tickets into draft help articles. Neither tells you which problems belong on the product roadmap, the onboarding checklist, or the sales team's agenda. That's the gap this prompt fills.

Run it every month. The first run usually surfaces three or four themes nobody had connected. The real value comes from running it again 90 days later and checking whether those themes have shrunk.

Prompt structure

Paste this into Breeze Assistant and replace the details in [CAPITAL LETTERS] with your own.

Role: You are a customer support operations analyst for [COMPANY NAME], a [INDUSTRY] business. You specialise in root-cause analysis: working out why customers need help in the first place.

Task: Review the tickets created in [TIME PERIOD, e.g. the last 90 days] in the [PIPELINE NAME] ticket pipeline. Group them into no more than 8 root-cause themes. For each theme, name the team best placed to stop these tickets arriving, and recommend one specific upstream fix.

Context:
- Our ticket category values are: [LIST YOUR CATEGORY VALUES]. Treat these as a starting point only. One category may hide several root causes.
- Teams that could own a fix: Product, Onboarding, Sales, Billing/Finance, Marketing, Support (documentation and knowledge base).
- Our customers are: [SEGMENT DESCRIPTION, e.g. "UK mid-market manufacturers, 50 to 500 employees"].
- Issues already being worked on (don't report these as new): [LIST, or "none"].

Constraints:
- Base every theme on the ticket subject, description, category, and source. Don't invent tickets, quotes, or figures.
- Show the ticket count for each theme and its share of total volume as a percentage.
- Only report a theme if it has at least [5] tickets. Group everything else under "Other".
- For each theme, say whether a knowledge base article would deflect it, or whether it needs a change elsewhere in the business.
- If the data is too thin to support a conclusion, say so instead of guessing.

Output format:
1. A table with these columns: Theme | Ticket count | % of total | Example ticket subjects (up to 3) | Likely root cause | Owner team | Deflectable or fixable | Recommended upstream fix
2. One short paragraph naming the single theme that would cut the most volume if fixed, and why.
3. A list of any tickets you couldn't confidently assign to a theme, with the reason.

Sample output

Sample output - illustrative, not from a real client portal. Based on a fictional B2B software company with 412 tickets over 90 days.

Biggest single win: fixing the quote template. It accounts for 17% of volume, and every one of those tickets starts with a customer who already feels misled about price.

Couldn't assign: 6 tickets with no description and the subject "Help".

Why this prompt works, and how to adapt it

The owner column does the heavy lifting. A list of ticket themes is interesting. A list where every theme has a named team and a single fix is something you can take into a meeting and assign.

It separates deflectable tickets from fixable ones. Some tickets genuinely need a help article. Many don't, and publishing an article about a broken quote template only teaches customers to read about your mistake. Asking Breeze to label each theme stops the knowledge base becoming the default answer to everything.

The 5-ticket threshold keeps one loud customer from setting the agenda. Without a minimum, a single account that raised 12 tickets in a bad week can look like a trend. Adjust the number to suit your volume: 5 works for a few hundred tickets a quarter, but you'll want 15 or 20 if you handle thousands.

The "already being worked on" line saves you from old news. If product already knows the sync is unreliable, you don't need Breeze to tell you again. Listing known issues pushes the analysis towards problems nobody has picked up yet.

Adapting it for your portal

Do you run more than one ticket pipeline? "Run this analysis separately for the [PIPELINE NAME] pipeline, then compare the top three themes across both pipelines."

Want to focus on your most valuable customers? "Only include tickets associated with companies where [TIER PROPERTY] is [Enterprise]."

Is your ticket category property patchy or unused? "Ignore the ticket category property and build themes from the subject and description only."

Briefing one team in particular? "Only show themes where the owner team is Product, and write each fix as a one-line roadmap request."

Is Breeze struggling with the volume? "Analyse the [100] most recent tickets only, and tell me which themes look likely to grow if the trend continues."

Beyond the prompt

Treat the output as an agenda. Take the table to a monthly meeting with the heads of product, onboarding, sales, and finance, and leave with one named owner and a date for each fixable theme. A theme without an owner will be back in next month's report.

To track progress properly, create a dropdown ticket property called "Root cause theme" using the themes Breeze found, and ask your support team to set it when they close a ticket. You can then report on theme volume over time in HubSpot, rather than relying on a fresh AI read each month.

Check Breeze's counts against a quick ticket report before you share anything. AI groupings are a strong first draft, and a two-minute sense check stops a miscount ending up in front of your leadership team.

Closing

Answering tickets quickly will always matter, but the cheapest ticket is the one that never gets raised. Once this is running monthly, success looks like a shrinking top theme every quarter, and a support team that spends more time on the questions only they can answer.

FAQs

Can Breeze Assistant analyse all my HubSpot tickets at once? Breeze Assistant can summarise ticket records and run counts and comparisons across your CRM data, but very large volumes may give you partial or less reliable results. If that happens, narrow the time period, analyse one pipeline at a time, or limit the prompt to your most recent 100 tickets. Always check Breeze's counts against a standard HubSpot ticket report before sharing them. 

Which HubSpot subscription do I need to run a ticket root-cause analysis with Breeze? You need tickets in your portal and access to Breeze Assistant. Some Breeze features use HubSpot Credits or need specific seats, and help desk features vary by Service Hub tier, so check what's included in your subscription before you start. 

How is this different from HubSpot's Knowledge Base agent? HubSpot's Knowledge Base agent focuses on turning resolved tickets into draft help articles. This prompt looks for themes that need fixing somewhere else in the business, such as a quote template, an onboarding step, or a product bug. It also labels which themes a help article would genuinely solve, so the two work well together. 

How often should I run a ticket root-cause analysis? Monthly works for most teams, looking back over the previous 90 days. That gives Breeze enough tickets to spot real patterns, and the overlap between runs shows you whether last month's fixes are reducing volume. 

What's the most common mistake with ticket theme analysis? Treating every theme as a knowledge base gap. Writing a help article for a problem caused by a broken process teaches customers to work around it, while the tickets keep coming. Give each fixable theme to the team that owns its root cause, with a named person and a date.