Blog
Notes on doing more with less.
Practical thinking on AI and automation: what actually works, and where it pays off.
-
What broke when we put an AI receptionist on a real phone number
We built one for ourselves and connected it to our own business line. Everything that went wrong went wrong after the browser testing passed, and most of it looked like the same problem until we pulled the logs apart.
-
AI receptionists for dental and allied health practices
The valuable call is not the new patient. It is the eleven o'clock ringing to cancel, and what happens in the ninety minutes after that decides whether the chair earns anything today.
-
AI receptionists for law firms: the first thirty seconds are the problem
By the time a caller has explained why they are ringing, they may have disclosed something you cannot un-hear and created a conflict you had no chance to check. That constraint shapes the entire design.
-
AI receptionists for real estate: five callers, one number
A buyer, a tenant, a landlord, a tradesperson and a potential vendor all ring the same line. Only one of them is worth interrupting an agent for, and telling them apart is the entire design problem.
-
AI receptionists for tradies: when every call is a triage call
A burst pipe at eleven at night and a bathroom quote for March arrive on the same number. Sorting one from the other is the whole job, and it is what most setups get wrong.
-
What software do you actually need to automate a trade business?
Most tradies asking this already own three quarters of the answer. The gap is almost never a missing app, it is the four layers not talking to each other.
-
Automating the day: scheduling, dispatch and the on-the-way text
A trade day is a sequence of promises about time, and almost all of them are kept or broken by hand. Here is which parts of that genuinely automate, and the one that changes how customers rate you.
-
AI receptionists for photographers and videographers
Every enquiry is a date question, and you cannot answer the phone with a camera in your hands or headphones on in the edit. That combination is what makes these trades different, and it changes what the system needs to do.
-
Do AI receptionists sound real? An honest answer
Mostly yes, and the voice stopped being the giveaway a while ago. What still catches callers out is the timing, and that matters more to your business than the accent does.
-
How much does an AI receptionist cost?
Nobody publishes one number because there isn't one. There are three lines on the quote, one of them is usually buried, and the fork between subscribing and building decides everything.
-
What should an AI receptionist ask your callers?
Most scripts are written from the business's point of view, which is why they collect the wrong things in the wrong order. Start from what the next person to touch the job actually needs.
-
What to look for in an AI receptionist: a buyer's checklist
Every product demos well. These are the capabilities that separate a system that books jobs from one that takes messages politely, and the questions vendors would rather you didn't ask.
-
How does an AI receptionist work?
Four steps, about a second and a half, and a lot of engineering in the gaps. Understanding the mechanism tells you exactly where these systems are strong and where they break.
-
Is an AI receptionist worth it? Work it out properly
The honest answer depends on numbers you probably haven't counted. Here's the arithmetic that settles it, including the cases where the answer is clearly no.
-
Will AI replace receptionists?
Mostly no, and the reasons are more interesting than the headline. What's actually happening is that calls nobody was answering are now getting answered.
-
From booking to invoice: joining up a service business
Support providers, allied health and mobile services lose most of their admin time in the handoffs between systems. Here's how the chain from booking to payment gets joined up.
-
AI guardrails for small business: a one-page policy
Large organisations write AI policies for months. A small business needs about a page. Here's what actually belongs on it, and what's safe to leave off.
-
Choosing an automation partner in Australia: what to ask
Anyone can connect two apps. The difference between a system that lasts and one you abandon in three months shows up in the questions asked before anything gets built.
-
Automating quotes and invoices without changing your accounting software
The slow part isn't the invoice itself, it's everything either side of it. Here's how to automate quoting and billing while keeping the accounting package you already know.
-
How to map a process before you automate it
The step that decides whether automation works, and it takes an afternoon. Here's the template I use, taken from sixteen years of doing this for government and higher education.
-
AI and automation for tradies: the four jobs to start with
You didn't start a trade business to do paperwork at nine at night. Here are the four admin jobs that automate cleanly for trades, ranked by what they actually give back.
-
AI receptionists: what they can and can't do
Missed calls are lost jobs, which is why AI phone answering has gone from novelty to something worth taking seriously. Here's an honest account of where it works and where it still falls over.
-
Zapier alternatives: n8n vs Make vs Zapier compared
Most people go looking for a Zapier alternative once the bill stops matching the value. Here's an honest comparison of the three that matter, and which one suits which kind of business.
-
An AI assistant that actually knows your business
Generic AI writes generic email, which is why most people give up on it. The difference between a novelty and a system that saves you a day a week comes down to what it's been given to read.
-
What AI automation actually costs a small business
Almost nobody publishes prices for this work, which makes it impossible to budget for. Here are mine, in Australian dollars, plus what moves the number and how to work out the payback yourself.
-
Why most small-business AI projects fail (and it isn't the technology)
The tools almost never cause the failure. Sixteen years of process work across government and higher education points to the same conclusion every time: the project was lost before anyone opened a laptop.
-
Three places AI quietly pays for itself
The highest-return AI projects usually aren't the flashy ones. They remove repetitive work you'd stopped noticing. Here are the three patterns that come up again and again, and how to tell if you've got one.