A phone can be answered at 2 a.m. even when no technician, driver, appointment, or dispatcher is available. That distinction should be designed into the call flow before a business advertises round-the-clock answering.
Separate five different promises
Document answering availability, information availability, message capture, human escalation, and actual service availability separately. A business may offer the first three at all hours without offering immediate field service.
Do the same for booking. Seeing an open calendar slot is not the same as confirming that the requested service, location, duration, equipment, or staff member fits that slot.
Write answers around current operations
Nova can state approved business hours, after-hours policies, and service boundaries. It should not transform “we respond to messages” into “a technician is on the way” or “we offer emergency work” into a guaranteed arrival time.
When service availability varies, use language that explains the next review step and who has authority to confirm it.
Build a human escalation ladder
Define which intents may attempt a transfer, who receives each one, during which hours, and what happens after decline, busy, timeout, or disconnect. Never route a transfer back to the same forwarded number without a tested loop-safe design.
If no human accepts, capture a complete record and use the approved voicemail or safe fallback. Avoid promising a callback time the business has not committed to.
Test the midnight version of the workflow
Call after hours as a new lead, an existing customer, an urgent caller, and a routine quote request. Verify what Nova says, whether the record is complete, and whether every ending is honest.
AfterTone can preserve the number customers already know and route only selected carrier-specific unanswered, busy, after-hours, or overflow conditions after configuration and testing with the customer. Production SMS remains on hold unless the separate gates are completed.