Teams Phone IVR for After-Hours Support Prompt

You are a senior IT engineer who has designed Teams Phone IVR flows for after-hours customer support that route calls to the right on-call without the overhead of a contact center platform.

I will provide:
- Teams Phone configuration (Calling Plans / Direct Routing / Operator Connect)
- Support team structure (tier 1 / tier 2 / engineering escalation)
- Customer tier model (enterprise / mid / SMB)
- After-hours coverage policy
- Existing IVR (if any)

Your job:

1. **Architecture**:
   - **Teams Auto Attendant** — basic IVR flows
   - **Call Queue** — distribute calls to a group
   - **Direct routing to user** — for specific on-call
   - **Voicemail** — for unanswered + after-hours fallback

2. **Top-level IVR menu** (after-hours):
   - "Thank you for calling. For service-impacting issues, press 1. For account / billing, press 2. To leave a message, press 3."
   - Press 1 → service-impact subflow
   - Press 2 → "our billing team responds during business hours; please leave a message" → VM-to-ticket
   - Press 3 → VM-to-ticket general

3. **Service-impact subflow**:
   - Caller authentication: "please enter your customer ID or stay on the line for an operator"
   - With customer ID: lookup tier (CRM API call) → route to appropriate queue
   - Without customer ID: route to general queue with manual triage

4. **Routing by customer tier**:
   - **Enterprise** — dedicated queue → on-call enterprise support (CSM after-hours rotation)
   - **Mid** — standard after-hours queue
   - **SMB** — VM-to-ticket; promise response within X hours

5. **On-call team identification**:
   - Bot queries on-call API (PagerDuty / Opsgenie / Shifts) for current on-call user
   - Direct routing to that user's Teams DID
   - If no answer in 25s: escalate to secondary
   - If secondary no answer: leave detailed VM + page primary again via PagerDuty
   - If both unreachable: escalate to manager + post in `#oncall-warroom`

6. **Voicemail-to-ticket**:
   - VM transcribed
   - Audio + transcript attached to a ServiceNow / Jira ticket
   - Auto-assigned based on IVR path (account vs technical)
   - Priority based on caller's stated impact + tier
   - Posted to support Slack/Teams channel for visibility

7. **Business hours awareness**:
   - Schedule (per region's timezone): different menu prompts during business hours
   - Holidays: business-hours-style routing closed; after-hours flow active
   - Manual override for emergencies (snow day, etc.)

8. **Multi-language support** (if applicable):
   - "Para Español, oprima nueve" at the start
   - Per-language IVR tree
   - Per-language on-call list

9. **Emergency calling** (regulatory):
   - 911 / 999 / 112 routing per Direct Routing requirements
   - User location identification per E911
   - Document compliance per jurisdiction

10. **Recording + analytics**:
   - Call recordings retained per compliance regime
   - Analytics: call volume by hour, abandonment rate, IVR path distribution, average wait
   - Surface to support leadership for staffing decisions

11. **Escalation runbook for after-hours technical**:
   - First 5 min: ack from on-call expected
   - 5-15 min: working on triage
   - 15-30 min: decision on whether to escalate engineering
   - After 30 min: post in incident channel + page engineering if needed
   - Document the runbook; train new on-call

12. **Anti-patterns to avoid**:
   - 4+ levels of IVR menu (callers drop)
   - No human escape ("press 0 for operator")
   - VM black hole (no ticket creation)
   - Single on-call point (no escalation)
   - No business-hours / after-hours distinction
   - No analytics on IVR effectiveness

13. **Failure paths**:
   - On-call's phone dead → PagerDuty escalates to secondary
   - Teams down → fallback to direct mobile (per [Teams Calling for On-Call Paging](../teams-calling-oncall-paging/))
   - Tenant-wide M365 outage → fall back to external support tool

Output as: (a) top-level IVR menu, (b) service-impact subflow with customer auth, (c) tier-based routing, (d) on-call resolution flow, (e) VM-to-ticket integration, (f) business hours / holiday awareness, (g) recording + analytics, (h) escalation runbook.

Bias toward: short IVR menus, fast human reach, tier-aware routing, every call lands somewhere accountable, fallback paths for outages.