Skip to main content
Twin is the personal AI operator in the Support domain. It reads the unified context, proposes next steps, and delegates to specialized agents so you can resolve faster and escalate less often.

​
Context Assembly

When you open a ticket, Twin scans the full projection: customer tier, linked Slack threads, recent Sentry errors, opened Linear issues, and related knowledge articles. It surfaces the five most relevant signals in the SwiftSolve panel so you do not need to search manually.

​
Proposals

  • Draft Replies: Twin drafts a response referencing the knowledge base and past similar tickets. You review, edit, and send.
  • Escalation Suggestions: If patterns match known engineering issues, Twin suggests escalation to Linear or Jira with a pre-filled description.
  • Classification Proposals: Twin labels new tickets by type, priority, and product area to speed triage.

​
Delegations

Twin hands off tasks to domain agents when specialized work is needed. For example, it delegates schema mapping for a new ticket type to Discovery, or routes a severe incident to the Incident Response agent for on-call coordination.

​
Continuity

Because Spine persists context, Twin resumes work across sessions. If you leave mid-incident and return tomorrow, Twin remembers the current status and proposes next steps immediately.