Five calls is the test Oncrew tells contractors to run before forwarding phones to an AI receptionist. The checklist says a night-time AI should do five things well: answer, collect contact and location details, classify the request, summarise the interaction, and alert the contractor team, while leaving pricing, scheduling, dispatch, ETA, site safety, and field decisions to humans. The scripted five-call exercise uses real-world scenarios contractors actually receive, from urgent-but-unclear faults to spam and vendor queries, and asks teams to read the AI summaries to judge whether the handoff is actionable. Contractors are told to use real calls rather than demo scripts and to review the summaries before they flip their after-hours lines to an automated system.

If an AI receptionist can't produce actionable handoffs, contractors can lose bookings and see missed-call revenue reach into the tens of thousands of dollars. That's the practical risk the five-call test is designed to expose, according to publishers and vendor guides that frame the problem. One guide states 62 percent of contractor calls go unanswered and that after-hours enquiries account for roughly 34 percent of booking opportunities. Vendor materials include client examples claiming reduced missed calls and increased bookings after deploying AI receptionists, but Oncrew’s checklist is about proving the system on the business’s own cases before any go-live.

What the five-call test covers

Oncrew published a structured checklist so contractors can judge whether an AI can take reliable intake without stepping into field decisions. The test asks teams to place five representative calls and then read the AI summaries to decide if the captured information is enough for a human to act.

Scenario one is an urgent but not fully diagnosable problem, like a ceiling leak, a breaker that keeps tripping, or a furnace that stopped overnight. In that case the AI should capture the caller’s name, callback number, address or service area, what changed, whether anyone is in immediate danger, and whether the caller wants an urgent callback. The checklist explicitly says the AI must not quote prices, promise a truck, or invent an arrival time.

Scenario two is a routine after-hours request for a quote, tune-up, inspection, or non-urgent repair. The AI’s job here is to classify the call as routine, capture the caller’s best callback window, and summarise the intent so the lead becomes a clean handoff for the next business day.

Scenario three covers existing customer follow-ups about jobs, invoices, parts, warranties, or return visits. Oncrew’s guidance tells teams the AI must gather enough context for the office to identify the customer and the issue.

If the system doesn't know an answer it should say so plainly and route the request for human follow-up.

Mastheads

Scenario four simulates commercial or property-manager calls. The system must identify who's calling, which property is involved, whether the caller is authorised, and what kind of issue is being reported. That information is intended to avoid mis-routed work orders and to protect contractors from taking instructions from the wrong contact.

Scenario five tests wrong-fit calls such as vendors, recruiters, and spam. The expectation is the AI recognises and routes these appropriately rather than treating them as routine leads that consume office time the next day.

The checklist’s underlying rule is simple and strict: automate intake, don't automate field decisions. Oncrew repeats that contractors retain ownership of pricing, scheduling, dispatch, ETA, site safety, appointments, CRM setup, and all field decisions. The recommended practical measure is to start evaluations with real call scenarios rather than demo scripts, so the system is judged by the cases the business actually sees after hours.

Australian vendors are already marketing AI receptionists built for contractor use cases, and their feature sets map closely to the five-call expectations. ADA Assist advertises fast setup, after-hours coverage, SMS and email summaries after each call, emergency alerts, and a 14-day free trial. ADA Assist lists a Starter plan at 199 AUD per month for after-hours coverage, a Pro plan at 349 AUD per month for 24/7 coverage and calendar booking, and a Business tier at 599 AUD per month.

CallMate positions itself as Australian-made, hosted on local data centres and offering a natural Australian voice. It advertises automatic appointment booking into Google Calendar, SMS confirmations, webhook and CRM integrations, and 24/7 coverage with no required credit card to start. Those features map directly to the checklist items that capture contact details, classification and summary, and then alert the office through SMS or webhook.

For larger rollouts some businesses prefer a done-for-you approach. Cybergarden describes an audit of existing phone lines and scheduling tools, followed by a three-week rollout that includes configuration and testing with platforms such as ServiceTitan and Jobber, and ongoing optimisation after launch. The pitch is that an integration service helps avoid common mistakes, like letting the automation make scheduling promises or writing arrival times into a CRM field without human verification.

Vendors and publisher guides position the five-call exercise as a readiness gate. The practical read is straightforward: the voice quality and the marketing demo matter less than whether the summary gives an office the context to assign the right job, follow up at a reasonable time, or escalate an emergency to a human. That's the narrow performance bar Oncrew sets.

Contractors evaluating after-hours AI receptionists are therefore instructed to run the five-call test with their own scenarios and to review the resulting summaries before forwarding their phone lines to any automated system. Doing the exercise on real calls surfaces gaps that polished demos can hide.

Related Articles

Before flipping after-hours lines to an automated receptionist, contractors should run the five-call test with real calls and review the AI summaries to make sure handoffs are actionable.

This article was created with AI assistance.