Skip to main content
Support work depends on clear, fast communication across humans, AI operators, and backend tools. The Communication layer in the Support domain defines how these participants exchange context and decisions.

Human to Human

  • Handoffs: When a ticket transfers between teams, Spine carries full context: history, Twin proposals, and open escalations. The next assignee starts with everything visible.
  • Escalation Notes: Engineers and support staff leave linked notes across Jira or Slack that appear in the Support workbench as part of the entity projection.
  • Shift Change: On-call responders inherit current incident status, active escalations, and pending approvals through Continuity.

Human to Twin

  • Natural Language Instructions: You override Twin proposals with a typed command, such as “draft a different reply focusing on downtime.”
  • Approval and Rejection: Approve or reject Twin drafts, escalation suggestions, and knowledge article proposals before execution.

Twin to Agent

  • Delegation Messages: Twin sends structured tasks to Incident Response, Escalation, or Knowledge agents with entity context and expected outputs.
  • Status Updates: Agents return progress and outcomes to Twin, which updates the workbench and notifies you if action is needed.

Agent to Agent

  • Discovery to Schema: When Discovery identifies a new custom field in Zendesk, it instructs Schema to update the entity model mapping.
  • Triage to Escalation: If Triage detects SLA breach risk, it signals Escalation agent to prepare an engineering issue.

Agent to Tool

  • PagerDuty: Incident Response pages on-call responders with severity, ticket links, and timeline.
  • Jira / Linear: Escalation Agent creates issues with reproduction steps and references to original tickets.