AI Lead-to-Booking Playbook for Service Businesses

An open calendar slot is not yet a valid service appointment: the job must also fit the service area, skill, equipment, duration, safety rules, and price authority. This playbook moves one request to a verified booking, safe route, or accepted human handoff only when those facts agree.

Copy the pilot card below into a document or spreadsheet and begin with one service request. If the catalog rule or dispatcher who receives an exception is unknown, stop and repair that gap first.

Copy the pilot card and define completion

• Start: one booking break has an owner, baseline, approved service catalog, coverage rules, and live capacity source.
• End: the request reaches a verified booking, accepted estimator/dispatcher handoff, safe alternative route, explicit follow-up state, or stop.
• Evidence: service/version, customer-stated need, location, permission, eligibility decision, booking ID, handoff receipt, correction, and outcome.
• Time horizon: one scheduling cycle including out-of-area, no-capacity, reschedule, cancellation, urgent/safety request, duplicate, and opt-out.

Roles and prerequisites

Map service code, scope and exclusions, service area, travel rules, skill/equipment, duration buffer, calendar, estimate/deposit authority, verified contact, consent, active job/complaint, and stop state. Publish emergency and out-of-scope routes before launch.

Phase 1: map request, resource, and booking state

• Owner: dispatcher with catalog owner.
• Inputs: catalog, coverage map, roster, calendar, CRM/job system, message events.
• Actions: trace inquiry through scope check, safety screen, resource match, booking, reminder, completion/change, and follow-up.
• Output: source/state map with owner, freshness, and precedence.
• Exit gate: every eligibility fact and resource has an authoritative source.
• Escalation/rollback: stop when calendar availability cannot be tied to the required skill/equipment or location.

Phase 2: select one service transition

• Owner: workflow owner.
• Inputs: journey evidence, baseline, common request types, reversibility.
• Actions: choose one narrow job such as standard inspection inquiry to appointment; define minimum questions, no-action rules, and human routes.
• Output: transition and acceptance criteria.
• Exit gate: service eligibility and safety can be decided from approved rules without diagnosing the underlying problem.
• Escalation/rollback: route ambiguous, hazardous, regulated, or custom-scope work.

Zalo describes an appointment-management utility for Official Accounts. It supports a booking mechanism, not service eligibility, capacity, quality, or outcome.

Phase 3: validate scope, safety, and capacity

• Owner: catalog owner for scope; dispatcher for capacity; safety owner for urgent signals.
• Inputs: customer-stated symptoms/need, location, current roster, equipment and policy.
• Actions: classify only against approved service rules; check coverage and resource bundle; route danger before commercial questions; avoid a firm estimate without authorized inputs.
• Output: eligible booking context or named route.
• Exit gate: scope, location, skill, equipment, time, and authority agree.
• Escalation/rollback: clear calendar holds and provide the approved emergency/out-of-scope path.

Phase 4: write, confirm, and reconcile the booking

• Owner: dispatcher/operations accepts exceptions and records the assigned service resource.
• Inputs: eligible context, approved message/action, live calendar and job events.
• Actions: create one idempotent booking; issue its ID only after success; revalidate before reminders; record changes and cancellations.
• Output: booked, changed, cancelled, routed, stopped, or human-owned state.
• Exit gate: calendar, job system, resource owner, and customer-facing states match.
• Escalation/rollback: suppress confirmation after partial write and restore the last verified state.

Google Business Profile documents customer-facing action links, including appointment links where supported. The source illustrates a discovery-to-action mechanism, not operational readiness.

Phase 5: review and expand by service cell

• Owner: workflow owner with dispatcher and safety owner.
• Inputs: valid bookings, reconciliation, scope conflicts, handoffs, cancellations, complaints, opt-outs.
• Actions: review by service code, area, channel, language, shift, and crew; inspect all safety and wrong-scope cases.
• Output: signed keep/pause/expand decision.
• Exit gate: quality thresholds hold through representative demand and exceptions.
• Escalation/rollback: disable only the affected service/area/channel cell.

Fictional completed run

A fictional home-maintenance company pilots standard electrical inspections in two districts. One customer reports a burning smell at the panel and asks for tomorrow; the safety rule stops booking and displays the approved emergency instruction and human contact without diagnosing the cause. Another customer requests a routine inspection, falls inside the service area, and matches an electrician/equipment slot. The system creates booking S-318, then sends confirmation. When the technician is reassigned, the reminder check detects the resource conflict and the dispatcher accepts the reschedule. No repair, safety, response-time, or revenue result is claimed.

Handoff, QA, and measurement

The Trigger column is the join key between the two compact tables.

QA checks: current catalog; exclusions; service area; minimum necessary questions; safety precedence; skill/equipment bundle; duration/buffer; price authority; duplicate booking; accepted handoff; reminder recheck; rollback.

The Metric column is the join key between the two compact tables.

Pause the affected service, area, or channel cell until correction is verified.

What to do after completion

Limits. This is not diagnostic, trade, safety, emergency, pricing, or legal advice. It claims no booking, response, service, or revenue benchmark, integration, or Easy AI capability.

Use the lead-response guide for response ownership, the appointment reminder template for a reconciled booking, and the AI-to-human handoff guide for exception context.