Categorize support tickets by issue type

A good set of categories comes from your own tickets, not from a generic template. Kite reads a sample of your recent tickets, including the customer's first message and the replies that followed, and groups them by the problem the customer actually had. It also checks the tags already in your helpdesk, so it can keep the ones your team relies on and avoid creating near-duplicates.

The categories Kite proposes are ones you can act on. "Shipping" is too broad to fix, because a late package, a wrong address, and a missing tracking link each have a different cause. "Order hasn't arrived," "Wrong item received," and "Can't find tracking number" each point to a specific fix. Kite also keeps the list short enough that your team can remember it and every ticket clearly belongs somewhere, and it includes a catch-all for tickets that fit nowhere else.

For example, a skincare brand selling on Shopify might end up with categories like these:

• Order hasn't arrived
• Returns and exchanges
• Damaged or leaking product
• Which product is right for my skin
• Discount code didn't work
• Change or cancel a subscription
• Skin reaction or ingredient concern

A software company using Zendesk would get a very different list, with categories such as login problems, billing questions, bug reports, setup help for a specific integration, and feature requests. Kite shows you the proposed list with a few example tickets under each category, and you can rename, merge, or split any of them before it starts tagging.

Kite reads each ticket's full conversation before it chooses a category. The subject line alone is often misleading. A ticket titled "Question" might turn out to be a refund request, and a ticket that opens as a sizing question might really be about an exchange. Reading the whole thread lets Kite tag what the customer needed, not just what they mentioned first.

Kite writes the category to the ticket itself as a tag. In Zendesk and Gorgias, it adds the tag to the ticket. In Intercom, it tags the ticket or the conversation, creating any new tags your list needs. Because the tags live in your helpdesk, your team's existing views, reports, and automation rules can use them right away.

When a ticket genuinely covers two problems, such as a late order that also needs a refund, Kite can apply both tags. When a ticket is too unclear to call, Kite puts it in the catch-all rather than guessing. If the catch-all starts filling up with similar tickets, that's a sign your list is missing a category, and Kite proposes adding one.

You can ask Kite to tag a first batch and check its calls before it works through the rest of your backlog. When you correct a call, Kite records the correction as a proposed rule. Once Kite's nightly review confirms it, the rule guides later tagging, so the same kind of ticket gets tagged the way you want next time.

Categorizing your backlog once tells you about the past. Keeping new tickets tagged tells you what's changing. Kite turns the tagging into a recurring job that runs on the schedule you agree on, such as every morning, and tags every ticket that came in since the last run.

The job updates your helpdesk directly, so it doesn't need to post anything. If you want a summary, Kite can send one to a Slack channel, email it to you from your team's Kite address, or keep each run as a conversation in the Kite web app. If your team needs tickets tagged within moments of arriving, your helpdesk can call a Kite webhook whenever a new ticket is created, and Kite tags that ticket when the call comes in.

Once tickets are tagged, Kite counts them. It shows how many tickets fell into each issue type, what share of your total volume each one represents, and how those numbers changed from the previous week or month. It also points out which issues are growing fastest, which is often more useful than which issue is biggest.

Numbers alone don't tell you why a problem is happening, so Kite reads across the tickets in each category and explains what they have in common. Say "Order hasn't arrived" made up 28 percent of your tickets last month, and most of them came within three days of purchase. Kite might notice that your product pages promise shipping in one to two days while your fulfillment usually takes four. That's a page problem, not a carrier problem, and it's the kind of finding that's hard to see when you're answering tickets one at a time.

Kite delivers this as a report with charts and real ticket quotes. If you want to keep checking it, Kite can build a dashboard at a stable link that refreshes its numbers when you open it. It also saves what it learns to your team's knowledge base as customer feedback, so later work, such as a new landing page or an email campaign, starts from the problems your customers actually have.

The point of categorizing tickets is to have fewer of them. Many recurring support questions exist because a page, email, or policy left the customer guessing. Kite works on the ones it can reach:

• It rewrites a product page or shipping policy so the promise matches what actually happens.
• It adds missing answers to your FAQ, the same way it writes a product FAQ from customer questions.
• In Intercom, it updates help center articles and the content Fin answers from when they describe an old price or feature.
• It drafts an order confirmation or onboarding email that answers the question before customers ask it.

Kite shows you each change before it goes live, and it publishes the ones you approve. Then it keeps checking the tagged ticket counts. If tickets about orders that haven't arrived drop from 28 percent to 12 percent after the shipping promise changes, you can see it. If the number doesn't move, Kite looks for a different cause and proposes the next fix.