Exception Path Design and Escalation Routing (Delivery)

Exception Path Design and Escalation Routing (Delivery)

1. 01

   Enumerate every branch where a workflow can stall, misfire, or produce a wrong-state record

   Walk the client's live journeys and list failure points: duplicate contact creation, scoring thresholds that never fire, SMS sends to opted-out numbers, and handoff tasks that land on inactive owners.

2. 02

   Assign each failure point a severity tier tied to client revenue exposure

   Tier 1 covers anything that touches billing, consent, or a booked meeting; Tier 2 covers nurture delays under 24 hours; Tier 3 covers cosmetic or reporting-only issues.

3. 03

   Define the detection signal for each tier before writing the fix

   A Tier 1 signal might be a webhook failure rate above 2% over 15 minutes; a Tier 3 signal might be a missing tag on a contact record. Detection precedes remediation.

4. 04

   Route Tier 1 alerts to a named human within 15 minutes, not to a queue

   Queues hide ownership. Name the on-call delivery lead in the escalation doc and confirm the client's own contact for anything that touches consent or payment.

5. 05

   Build the fallback action for each automated branch

   When a scoring model returns null or a chat flow drops mid-conversation, the fallback should create a manual task with the contact context attached rather than silently ending the journey.

6. 06

   Log every exception with timestamp, journey name, contact ID, and resolution

   This log becomes the evidence base for the monthly retainer review and for scoping the next optimization sprint.

7. 07

   Rehearse the escalation path with the client's team once before go-live

   A 30-minute tabletop run through two Tier 1 scenarios surfaces disagreements about who calls whom far cheaper than a live incident.