Support Ticket Pattern Finder
Count recurring support-ticket symptoms from a redacted CSV, using rules you can inspect and change. Get a traceable report to review with your team.
For support leads, CX teams and operations teams reviewing repeated ticket subjects. Literal matching keeps the counting reproducible; people decide what the matches mean.
SYNTHETIC DEMO / ACTUAL ENGINE OUTPUT
13 input rows · 1 exact duplicate excluded
Evidence: T-001, T-002, T-005, T-009
Evidence: T-003, T-006, T-010
Evidence: T-004, T-007, T-011
Evidence: T-008, T-012
Each share uses the 12 eligible tickets as its denominator. Review the subjects behind the IDs before deciding what to fix.
Synthetic tickets only. No customer data or measured business outcomes.
From ticket subjects to counts you can inspect.
Redact your CSV
Use pseudonymous IDs, ISO dates and redacted subjects. No helpdesk connection or credentials are required by the engine.
Apply explicit rules
Choose literal match-all and match-any terms. The engine accounts for duplicates, conflicting IDs and excluded rows.
Inspect the evidence
Read counts, eligible-ticket shares and source IDs in JSON or Markdown. Check exclusions and Other/Unclear before drafting your own follow-up.
Instructions, executable logic and examples.
Version 0.2.0 includes the working engine and its supporting material. Public download remains pending release review.
- SKILL.md and one Python analysis engine
- Default and custom rule examples
- Synthetic CSV, JSON and Markdown reports
- Install, update, removal and recovery guides
- Local HTML help dashboard and usage terms
Keep sensitive tickets out of the preview.
This page has no ticket upload or product email form. The standalone engine makes no network requests, does not connect to a helpdesk and does not send customer messages. Reports can still contain sensitive IDs or custom theme names: redact inputs and review outputs before sharing.
An AI client may process anything you give it. Local engine behavior does not change your client's privacy settings or policies.
Preview now. Downloads are not open yet.
The engine and extracted package have passed automated tests with synthetic data. Independent fresh-install and native-client verification are still pending, so this is not a verified Beta or production release.
There is no checkout or product download on this page. The sample above lets you judge the output while release checks continue.
Review the sampleBefore you try it
What input does it accept?
UTF-8 CSV with exactly ticket_id, created_at and subject columns. Dates use YYYY-MM-DD. Use authorized, redacted subjects and pseudonymous IDs. Limits: 2,000,000 bytes, 10,000 data rows and 4,000 normalized characters per subject. XLSX and free-form text are not supported.
How does the matching work?
Configurable literal match-all and match-any terms group ticket subjects. Ambiguous and unmatched tickets go to Other/Unclear. Exact duplicates count once; conflicting IDs are excluded before optional inclusive date filtering.
Does it infer root causes?
No. This is literal phrase matching, not semantic inference. It does not interpret negation, sentiment or resolved status. A match does not prove an issue exists. People review the evidence and decide what to do.
Where does ticket data go?
The standalone Python engine makes no network requests and writes reports to standard output. This preview accepts no ticket uploads. An AI client may process anything you give it; review that client's data handling before providing data.
Can I download it now?
Downloads are not open yet. Version 0.2.0 has been exercised with synthetic data, but independent fresh-install and native-client release checks are pending. This is a preview, not a verified Beta or production release.
What does it require?
The engine requires Python 3.10 or later and uses the standard library. An AI client may run the Skill subject to its permissions. ChatGPT Work and Claude fresh-install compatibility has not yet been independently verified.