Playbook

The Missed Call Leak: What an Unanswered Phone Really Costs

By David Saleh · Silas Main · 17 July 2026

Second ring, third ring. A plumber's hand is inside a wall cavity holding a compression fitting steady, and the phone goes to voicemail on the fourth. Whoever called doesn't leave a message. They call the next name on the list.

Nobody wrote that lost job down anywhere. That's what makes the missed call leak the most expensive problem most small businesses have never measured.

What the phone actually costs

Do this with your own numbers, not ours. Take your average job or booking value. Multiply by how many calls you'd guess go unanswered or unreturned in a normal week, the ones where nobody left a voicemail, or left one that never got a callback before the customer moved on. Multiply by 52.

Most owners have never run that number, because you can't count what you didn't hear ring. Once you do run it, most small businesses land somewhere between "concerning" and "that explains a lot about this quarter."

The same leak, five different industries

The mechanics don't change much from one small business to the next. Only the shape of the loss does.

A trades business loses the emergency call: the burst pipe, the no hot water, the job that would have paid same-day rates if someone had picked up.

A tour operator loses the impulse booking: someone scrolling a phone on a Friday night, deciding tonight, moving to the next operator's site the second yours doesn't answer.

A care provider loses trust before they lose the booking: a family reaching out about something sensitive, going quiet when nobody replies for two days, and not reaching out again.

An event business loses the big-ticket enquiry: a wedding, a corporate function, a booking worth months of smaller jobs combined, sitting in a form submission nobody saw over a weekend.

A kids' activity studio loses the parent mid-decision: comparing three studios in one evening, booking whichever one texts back before bedtime.

Different industries, same shape. The enquiry arrived. Somebody wanted to give you money. The business found out about the loss, if it found out at all, weeks later, as a quieter month than it should have been.

Why "we'll call them back" doesn't work

Voicemail assumes two things that are usually false: that the caller leaves a message, and that whoever calls back reaches them before they've already booked someone else. Neither holds for long in a trade that competes on speed as much as price.

The fix isn't a bigger phone, or a receptionist nobody can afford to hire full time. It's a system that reads and sorts what comes in the moment it lands, so nothing sits in a queue nobody's watching.

Inside a real build: a plumbing office in Adelaide

We built an enquiry triage system for a plumbing company here in Adelaide. It checks the office inbox on a short cycle: missed-call notes, website form submissions, direct emails. Each one gets sorted by urgency, from a genuine emergency down to routine supplier mail, and a reply gets drafted in the office manager's own signature, ready for her to read, edit if she wants, and send.

It doesn't send anything itself. If something looks like a genuine emergency, it raises an internal alert straight away instead of waiting in line with everything else. And it knows what it doesn't know: if an email mentions an attachment the system can't read, it never claims that attachment is missing. Guessing there's a problem when there isn't one is its own kind of expensive mistake, and it's one we build out on purpose. The full six-way sort we use, and the exact prompt behind it, is in the enquiry triage prompt linked below.

Inside its first week live, this system caught a genuine emergency that would otherwise have sat in an inbox overnight. The office manager's reply time on ordinary jobs went from a five minute typed email to reading a draft and clicking send.

What "answers, sorts, and books" actually means for calls

The same idea extends to the phone itself: something picks up, asks the two or three questions that matter (what's needed, when, contact details), and books straight into a calendar if the answer's clear enough. We run a version of this on our own studio line, for exactly the same reason: it's the same leak, just on the phone instead of in an inbox.

None of this replaces a person deciding what to do about an emergency, a tricky customer, or a job that needs judgement. It replaces the part where a customer's enquiry sits unread for six hours because everyone's on a job.

The guardrails that keep it safe

A system like this only earns its place if it's boring in the right ways.

It drafts. It doesn't send to a customer, ever, without a person reading it first.

It flags. Anything that looks like a genuine emergency gets raised immediately and loudly to a person, not filed quietly for later.

It defers. If it doesn't know something, it says so instead of guessing, and that's tested deliberately before anything goes near a real customer. Exactly how that test works is in the chatbot grounding playbook, linked below.

Start with the leak, not the tech

Before any prompt, any AI, any new tool: work out where your business is actually losing enquiries right now. The prompt below does that in about the time it takes to make a coffee.

The prompts

You are helping me find out what missed enquiries are actually costing [BUSINESS NAME], a [INDUSTRY] business in [SUBURB OR CITY].

Here's what I know:
- Average job or booking value: $[X]
- In a normal week, we estimate we miss, don't answer in time, or lose track of about [NUMBER] enquiries. Count phone calls that went unanswered, forms nobody replied to same day, and DMs that sat unread.
- Our current process when someone reaches out: [describe, e.g. "rings out to voicemail", "goes to a shared inbox", "DMs sit until someone checks socials"].
- Of the enquiries we DO answer quickly, roughly [PERCENT]% turn into a booked job.

Using only this information:
1. Calculate the estimated number of enquiries lost per year, and the estimated dollar cost, showing your working.
2. Compare that cost to what a basic triage or answering system might reasonably cost to run in a month. If you don't have a real figure, ask me for one rather than inventing it.
3. Identify the single weakest point in how enquiries currently reach us, the channel most likely to be missed.
4. Suggest the smallest possible first fix for that one weak point, something a person could set up or approve this week, not a full system.

Keep it under 350 words, plain English, and show the maths rather than just the conclusion.

You are helping me audit what happens after hours at [BUSINESS NAME], a [INDUSTRY] business in [SUBURB OR CITY].

Our current hours are [OPEN HOURS]. Outside of those hours, here's what actually happens when someone tries to reach us:
PHONE: [e.g. "rings out with no voicemail", "voicemail, checked once a day"]
WEBSITE FORM OR DM: [e.g. "sits until the next morning", "nobody's assigned to check it"]
WALK-IN OR AFTER-HOURS EMERGENCY: [describe what a customer sees or hears if they try to reach you right now, at 9pm on a Saturday]

Using only this information:
1. Estimate what proportion of our total enquiries likely arrive outside business hours, based on the nature of our business (a trade with emergencies will differ from a business that only operates during set hours).
2. For each channel above, say plainly what a customer experiences right now if they reach out after hours, and whether that reads as "we've got this" or "you're on your own."
3. Suggest the minimum viable after-hours coverage for this business: not a full AI receptionist necessarily, just the smallest change that stops a genuine emergency or a hot lead going cold overnight.
4. Flag anything after-hours that should still always reach a real person immediately, no matter what, and why.

Under 350 words, plain English.

Questions

Adelaide · ACST · ABN 17 626 827 646