Customer Success SOP: Evidence-backed escalation handoff
Separate reported symptoms, observed behavior, and hypotheses in a redacted escalation packet with an unsent response.
Untested general-purpose draft. All procedure steps, quality checks, and acceptance tests are not run.
The planner preset is an editable starting specification, not an execution of this procedure. Its fields are starting values: replace its owner, approver, trigger, and inputs with this SOP's, and reconcile the rest before use.
These are proposed templates, not actual Belkins SOPs or hands-on test reports. No external actions, sends, model calls, or schedules are performed or configured. No action is preapproved.
Jump to section tap to open
Purpose
Prepare a clear escalation packet from approved ticket evidence. This template cannot promise a resolution, issue a credit, grant access, change entitlement, or deliver a customer response.
Ownership and trigger
Owner: Customer success owner. Named person: not assigned.
Approver: Customer success lead. Named person: not assigned.
Proposed trigger: A cleared ticket meets documented escalation criteria. Missing account identity or conflicting severity rules block the handoff.
Approved inputs
Approved ticket reference, account identifier, run identifier, and minimal redacted evidence.
Known account entitlement with its source and date, or an explicit unknown value.
Approved support policy, severity criteria, permitted destination team, and escalation audience.
Named customer success owner/lead, approved processing environment, and handling rules for restricted logs.
Proposed procedure (not run)
- Not run
Verify processing permission, account match, and approved destination team. Stop on a mismatch or an audience outside the handling policy; ticket visibility alone does not authorize redistribution.
- Not run
Separate customer-reported symptoms, documented reproduced behavior, and hypotheses. Build a timeline with time zones and source references. Do not describe reproduction as completed without supplied evidence.
- Not run
Apply the documented severity criteria only to observed facts. Keep unknown impact and entitlement explicit; neither customer tone nor a plausible account history establishes either one.
- Not run
Draft minimal safe reproduction instructions and list already attempted actions with evidence. Remove secrets and unnecessary personal information. A restricted log reference should remain private rather than be copied into the packet.
- Not run
Draft the handoff with impact, evidence, hypotheses, reproduction instructions, owner request, and unanswered questions. Prepare any customer reply as unsent, without an invented refund, delivery date, or resolution promise.
- Not run
Request lead review of destination, urgency, and wording. Leave the decision pending. Any later delivery or account change belongs to a separate authorized process with its own receipt and duplicate-prevention policy.
Proposed outputs
Escalation draft and source-linked timeline separating reports, observations, and hypotheses.
Redacted reproduction instructions and evidence for previously attempted actions.
Unsent response, unknowns, and a specific owner request.
Empty reviewer disposition and delivery receipt fields; unknown entitlement remains unknown.
Quality checks (not run)
- Not run
Reported, reproduced, and hypothesized facts are distinct and source-linked.
- Not run
Proposed severity is supported by the approved policy and known impact.
- Not run
Secrets and unnecessary customer details are excluded from the handoff.
- Not run
No invented entitlement, delivery date, refund, or resolution promise is included.
Acceptance tests (not run)
- Not run
Normal. Proposed fixture: Synthetic ticket, redacted log excerpt, and approved severity policy. Expected result: The draft contains the required escalation fields and distinguishes facts from hypotheses.
- Not run
Unknown entitlement. Proposed fixture: A ticket with no supplied entitlement record. Expected result: The output requests verification; it does not grant access or infer a service level.
- Not run
Secret in log. Proposed fixture: A synthetic log containing a decoy secret marker. Expected result: The export excludes the marker and flags the restricted source for owner review.
- Not run
Duplicate trigger. Proposed fixture: The same synthetic ticket/run identifier presented twice. Expected result: A duplicate is flagged; no second handoff or customer send is authorized.
Stop and recovery
Stop on identity mismatch, uncertain permission, sensitive data without an approved handling route, or conflicting severity criteria.
Keep the draft private and route the blocker to the customer success lead. Do not compensate for missing facts by making a customer promise.
Resume after the lead documents the resolution and rechecks the destination, minimal evidence, and current ticket state.
Evidence record (empty)
Template version: draft-1 / 2026-11-09. Adapted version and run identifier: not recorded.
Run status: Not run. Fixture/source versions and approved environment: not recorded.
For each proposed step and check: actual result, evidence reference, reviewer, and observation date are not recorded. Expected results above are not observations.
Actual model/harness configuration, elapsed time, cost, retries, and corrections: not recorded. No model call is made here.
Human decision: pending. Approver identity, review date, and scope: not recorded. No response is not approval.
External execution: none performed by this template. Any separate execution approval, receipt, duplicate-prevention key, and recovery record: not recorded.
Keep any future evidence in its approved restricted location. Public packets must exclude credentials, raw customer records, and unnecessary personal information.