01
How does AI support ticket classification work?
A classifier reads the ticket text and predicts one of your support categories, such as Billing, Account Access, or Technical Support. StayCharted learns from labeled tickets and tests on held-out examples. Your support system uses the returned label and confidence to route or request review.
02
Your support team already has the training data
Most support organizations have years of examples showing how requests should be handled. A customer writes: “I'm unable to add another user to our account.” Your team may categorize that as Account Administration and route it to a particular group. Another customer asks: “Why was I charged twice this month?” That might be Billing.
People on the support team learn these distinctions over time. Generic AI does not necessarily know them because every company organizes support differently. StayCharted AMT lets you teach AI using the decisions your team has already made. AI Classifier Models use natural language processing to understand ticket text and learn your support categories.
03
Start with a spreadsheet
Export historical tickets from your support system with columns such as:
| Ticket Text | Category |
|---|---|
| Cannot reset my password | Account Access |
| Invoice contains duplicate charge | Billing |
| Mobile app crashes after login | Technical Support |
Upload the file to StayCharted AMT and select the text and label columns. StayCharted AMT trains a model specifically around those examples. You can then test new tickets against the model and review its predictions before using it operationally.
04
Your categories remain yours
You don't have to reorganize your business around a generic AI taxonomy. If your company uses:
- Billing
- Account Access
- Product Issue
- Feature Request
- Cancellation
- Partner Support
the model learns those categories. If another company uses completely different categories, its model learns those instead.
05
Move from spreadsheets to automation
For smaller teams, periodically exporting tickets and processing them in StayCharted AMT may be enough. As volume grows, the same model can be connected to your application through the StayCharted AMT Prediction API.
- New Support Ticket
- Your Support Application
- StayCharted AMT Prediction API
- Category + Confidence
- Route to the Right Queue
The AI does not need to replace your support platform. It becomes one small decision-making component inside the workflow you already use.
06
Start with a decision your team makes every day
If people repeatedly read a short piece of text and decide which category it belongs to, you may already have a good candidate for StayCharted AMT.
07
What is automated ticket triage?
Automated ticket triage assigns incoming requests to categories or queues. StayCharted AMT learns these assignments from previously handled tickets; your application decides whether to route or request review.
08
Can I use my own support categories?
Yes. Train on tickets paired with your team's approved categories, then test how the model assigns new requests.
09
Does StayCharted replace my helpdesk?
No. You can export tickets to a spreadsheet or build an API integration on a supported plan. StayCharted returns categories and confidence; your helpdesk owns assignments, notifications, and resolution. A native integration with a particular helpdesk should not be assumed.
10
How do I know whether ticket classification is accurate enough?
Inspect held-out accuracy by category and the mistakes your team would need to correct. Test on representative new tickets, including mixed requests and rare categories. Choose review rules for your workflow; a public banking benchmark does not establish accuracy on your own support data.