// Service dashboard — SLA %, MTTR, CSAT, trends · demo accounts available on request
The problem → the approach
Service desks drown in mis-routed tickets, silent SLA breaches and duplicate reports during outages, while the knowledge that would have deflected half the queue goes unused.
Relay runs the desk inside Postgres: tickets are classified and routed on arrival, SLAs follow business hours and pause while waiting on the customer, automations and approvals run on triggers, and Relay Assist drafts the reply.
What it does
Relay Assist (real AI)
pgvector + gte-small embeddings for triage, similar tickets and article suggestions; summaries and reply drafts.
Business-hours SLAs
First-response and resolution targets per priority, paused while pending, with a pg_cron breach watchdog.
Routing & automations
Category → team with round-robin or load-balanced assignment, plus a JSON rule builder that runs on triggers.
Approvals & catalog
No-code catalog forms, manager approvals, CAB for changes and fulfilment tasks that auto-resolve the request.
Major incidents & problems
Status-board banner with 'I'm affected too'; resolving a problem resolves every linked incident.
Dashboard & realtime
SLA %, MTTR, CSAT, arrival heatmap and agent leaderboard; collision detection shows who's typing.
Screens
Who uses it
Security, enforced in the database
SECURITY DEFINER functions that verify the caller before changing anything.