1. Cancellation-fee source card
Copy/paste fields:
- Customer/job, appointment date, service type, and CRM/job number:
- Request type: cancel / reschedule / fee dispute / refund request / credit request / no-show complaint / policy question:
- Verified source: booking terms, signed estimate, invoice, text thread, call note, website policy, manager note, dispatch record, technician arrival log, or payment record:
- Customer reason exactly as stated:
- Current fee/charge/credit status: not charged / pending / invoiced / paid / disputed / refund requested / unclear:
- Policy language we can quote safely:
- Exception request, if any:
- Approved owner/manager reviewer:
- Customer-safe next step: confirm policy / review exception / reschedule / issue credit / deny waiver / hold for owner review:
- CRM labels: cancellation_fee_review / reschedule_exception / fee_dispute / owner_review_needed / refund_review / policy_question:
- Customer update due time and next check date:
2. Customer-safe reply snippets
Reviewing before a decision: “Hi [Name], I see your note about [cancellation/reschedule/fee]. I’m going to review the appointment record and our policy before we promise a waiver, credit, or refund. We’ll update you by [date/time].”
Approved reschedule without fee promise: “We can look at the next available appointment window. I’m marking the fee question for manager review separately so scheduling does not accidentally create a refund or waiver promise.”
Policy explanation: “Our record shows [verified policy/source]. Based on that, the current status is [fee/credit/refund status]. If you want us to review an exception, I can route it to [owner/manager] with your reason included.”
Do not say: “we will waive it,” “you qualify for a refund,” “the fee never applies,” “we can guarantee the earlier slot,” “the technician was wrong,” or “policy requires this” unless the approved source and reviewer support it.
3. AI draft prompt with guardrails
You are helping a local-service business draft a customer-safe reply about a cancellation, reschedule, no-show fee, fee waiver, credit, or refund request.
Use only the verified facts below. Do not invent policy terms, refunds, waivers, credits, legal rights, payment status, technician fault, appointment availability, warranty coverage, customer intent, or owner approval.
Facts I can verify: [paste cancellation-fee source card]
Create:
1) a short internal owner/manager summary;
2) one customer-safe reply under 650 characters;
3) missing facts/questions that must be resolved before promising a fee waiver, refund, credit, charge, reschedule priority, or policy decision.
If the policy or approval is unclear, write NEEDS OWNER REVIEW instead of a promise.
STOP AUTOMATION
- Do not let AI approve, deny, refund, waive, charge, credit, or reschedule based on incomplete notes.
- Stop if the message involves legal/policy interpretation, card disputes, chargebacks, refunds, credits, recurring billing, technician fault, safety, warranty coverage, or a high-value customer relationship.
- Do not invent appointment history, consent, notice timing, arrival attempts, payment status, or exception approval to make the reply sound complete.
- Escalate to an owner or manager when the policy source, customer reason, dispatch record, and payment record conflict.
4. Weekly review questions
- Which cancellation or no-show fees were disputed without a policy/source link in the CRM?
- Which exceptions were granted repeatedly: weather, illness, technician delay, access issue, late part, dispatch error, or relationship save?
- Which customer-safe wording reduced repeat calls without promising refunds or credits?
- Should Local Lead Rescue add a cancellation-fee review lane with policy source, exception owner, payment status, customer update cadence, and STOP AUTOMATION rules?
Product bridge: Pair this with Local Lead Rescue appointment-confirmation, cancellation/reschedule, service-call fee explanation, chargeback evidence, and refund-exception resources so policy-sensitive replies are tracked before AI drafts customer messages.