Shelter Surge Playbook

Resources → Add (or edit existing). Set total capacity, current occupancy, services available (beds, meals, hygiene, medical, pet-friendly), and hours. For surge response, also fill in 'overflow plan' in the notes — where do people go if you fill up? AI matching surfaces this to the coordinator when capacity is tight.

• Total capacity reflects fire-marshal limit, not aspirational number
• Services accurately listed (don't list 'medical' unless someone qualified is on-site)
• Overflow plan documented in resource notes

During surge, update occupancy every time it changes by more than 5%. Use the mobile capacity edit — tap the number, adjust, save. Set a 1-hour stale-capacity reminder during active surge (vs. the default 4 hours). If a status change is dramatic (full → emergency overflow), add a note so other coordinators see context.

• Capacity owner on shift with mobile access
• Stale-capacity reminder dialed down to 1 hour for surge
• Status notes added at each major change

Decide who handles overflow before you need to. Settings → Partners → add the partner shelters that will take spillover, with each one's intake contact and channel. Inside your shelter resource, set 'overflow partner' to your primary spillover destination. When you hit amber/red status, the triage engine suggests the overflow partner automatically.

• At least two overflow partners added (primary + backup)
• Each partner contact has been notified they're in your overflow plan
• Overflow partner linked from your resource record

When capacity hits 90%, flip status to amber and start routing the next wave to your overflow partner. At 100%, flip to red so the matcher stops suggesting you entirely. Use 'refer' (not 'route') for handoffs to a partner org so the partner is in the audit trail. Track every refusal so the after-action report shows true unmet need.

• Amber at 90%, red at 100% — no aspirational over-promise
• Every overflow case uses 'refer' so partner ownership is recorded
• Refusals logged with reason (not just closed silently)

After the surge, run a resource utilization report for the period. Look at peak occupancy, total served, refusals, and average dwell time. Capture what changed in capacity owner's workflow — if anyone struggled with the mobile capacity edit, schedule a re-train. Reset stale-capacity reminders to the default 4 hours so you're not getting noise during steady state.

• Utilization report saved
• Capacity owner workflow reviewed, any friction noted
• Stale-capacity thresholds returned to steady-state defaults