Support Triage Model

Every support case should produce a category, severity, evidence checklist, customer-safe response, and escalation decision. The AI assistant may draft these fields, but staff remain responsible for review.

Classify first

Classify the case by product area: recordings, meetings, captions, translation, billing, workspace access, custom domains, SSO/SCIM, security, or general support. Use the customer’s words plus timeline and diagnostics evidence.

Establish impact

Capture the affected workspace, requester email, feature, affected object ID, customer-visible error, first seen time, whether the issue is ongoing, and how many users are affected.

Choose severity

• Critical: account takeover concern, data exposure concern, owner lockout for a business-critical workspace, billing outage affecting many customers, production-wide recording or meeting outage.
• High: one tenant blocked from a paid critical workflow, failed owner recovery, repeated payment/entitlement mismatch, custom domain takeover concern.
• Medium: isolated export failure, one meeting caption issue, single user access problem with workaround.
• Low: how-to question, plan explanation, cosmetic issue, or non-urgent request.

Support response rule

The customer reply should state what is known, what is being checked, and what information is needed. Do not promise a fix, refund, transfer, or data recovery until the guarded workflow has completed.

Escalate when

Escalate if sensitive data may be exposed, identity authority is unclear, a high-risk action is requested, a paid workspace is blocked without workaround, or diagnostics show repeated platform failure.