The 11pm call. Capture it, do not try to close it.

The actual problem with after-hours

Not that you miss calls. That you cannot triage them without answering them, so the choice is between answering everything at 11pm or missing the one that mattered.

A voicemail solves neither. Most people do not leave one, and the ones who do leave a name and a number and no indication whether the ceiling is currently coming down.

What a good after-hours agent does

Answers, works out whether this is urgent, captures the details that make a callback productive, and tells the caller honestly what happens next and when.

That last part is most of the value. A caller who knows they will hear back by 7am usually stops ringing around. A caller who reaches a machine that says nothing definite rings the next number.

• Answer in a way that does not pretend It should not claim to be a person. Callers work it out and resent the pretence more than the machine.
• Establish urgency Water, gas, no power, someone unsafe. Four questions sorts genuine emergency from Monday.
• Capture the callback essentials Address, access, what is happening, best number. Enough that the morning call is productive.
• Commit to a time And a real one. "Someone will call before 7am" beats "we will be in touch".
• Escalate the genuine emergency To an actual phone, immediately. That is the whole reason it is worth building.

What it should not do

Quote. Book. Diagnose. All three demo well and all three fail badly, because a wrong price given at 11pm is a price you will spend a week walking back.

The failure is asymmetric and that is why the rule is firm: capture-and-callback is boring and works; close-the-job is impressive and occasionally costs you a customer plus a reputation.

This is one of the narrow genuine AI cases

Unstructured input, real judgement about urgency, high enough volume to matter, and a mistake that is visible and recoverable. That combination is the test for whether an AI step earns its cost, and after-hours passes it — unlike most of what gets sold under the same heading, per the AI tools page.

It is also a case where a rule cannot substitute. You cannot write an if-statement that distinguishes a burst pipe from a dripping tap in a caller’s own words.

Where this does not pay

If you get one after-hours call a fortnight, the running cost is not justified and a well-worded voicemail with a clear callback promise does most of the job.

If you do not actually want after-hours work, do not build this. Being reachable creates an expectation of being responsive, and an agent that captures calls you have no intention of returning is worse than not answering.

And if the calls are mostly existing customers asking where their job is up to, the fix is upstream. That is a status-visibility problem, and answering it better at 11pm treats the symptom — getting the systems to talk removes the question instead.

Before building anything here, count the calls for a fortnight. Most people over-estimate their after-hours volume because the memorable ones are the 2am emergencies, and the arithmetic looks different once you have a real count.

What we have not measured, and what we would measure for you

There is no reliable published Australian figure for how long after-hours call handling takes in a small business, and we are not going to borrow one from a vendor whitepaper to make this page more persuasive. Most numbers you will see quoted for this were produced by companies selling the fix.

What we would do instead is count it on one of your jobs. Follow a single job end to end, note every time a human touches this step, and multiply by what that person costs loaded. That is a real number about your business and it takes an afternoon.

You can run the arithmetic yourself first — the leak calculator uses ABS, Average Weekly Earnings, Australia, May 2026 for the wage default and ATO, Super guarantee, current rate for super, and returns a range rather than a single confident figure. If the range comes back under our minimum engagement, that is a real answer and you should stop there.