Automation · Getting started

What Should I Automate First? A No-Hype Guide for Small Business Owners

Forget the robotic flyers and generic ChatGPT copy. The real value of AI is in the work your customers never see: faster follow-up, fewer missed opportunities and less repetitive admin.

Forget the robotic flyers and generic ChatGPT copy. The real value of AI is in the work your customers never see: faster follow-up, fewer missed opportunities and less repetitive admin.

If your experience of business AI has mostly involved strangely polished Facebook flyers, generic blog posts and emails that begin with "I hope this message finds you well," you'd be forgiven for wondering what all the fuss is about.

Small business owners have been told that AI will transform everything. Then they try it, spend twenty minutes arguing with a chatbot, receive something that sounds vaguely American and decide it would have been quicker to do the job themselves.

That scepticism is healthy.

Most of what people currently recognise as AI is content generation: tools producing words, pictures, videos and increasingly questionable social posts. It's highly visible, easy to experiment with and often not particularly useful.

But that isn't where AI delivers its greatest value to most small businesses.

Placeholder image
Placeholder — final image to come.

The useful AI is often the part nobody sees.

It's the system that responds to a missed call before the customer rings your competitor. It's the new enquiry that arrives in your CRM already qualified and researched. It's the appointment reminder that goes out without somebody remembering to send it. It's the follow-up that happens on time, every time, even when your team is flat out.

Much of this is not new. Businesses have used automation for decades. What AI changes is the range of information those automations can understand — and therefore how much useful work they can handle.

The opportunity isn't to make your business look more "AI." It's to make your business work better — and that starts with knowing which piece of the business to fix first.

12/wk
Missed-call leads recovered for Highlands Cleaning & Maintenance with text-back automation
~$180k
Recovered by Sydney Solar & Battery Company from 1,000+ dormant leads
30 → 5 min
Owner research time per enquiry, Simply Draft, once lead intelligence ran automatically
47 → 5 min
Average first response time, Resolved IT, after connecting ticketing and triage

The four places automation earns its keep first

Every business is different, but in practice the same four gaps show up again and again in the businesses we work with. They're worth checking in this order, because each one is closer to money already walking out the door than the last.

1. The moment a call goes unanswered

A missed call is the clearest, most measurable leak in most small businesses — and the easiest one to fix. If nobody picks up, the caller doesn't wait around; they try the next name on the list. An automatic text-back within seconds of a missed call — acknowledging the enquiry, asking what's needed, offering a real booking slot — turns a lost call back into a live conversation before the caller has hung up on the idea of using you at all.

Highlands Cleaning & Maintenance now recovers around 12 missed-call leads a week this way. It's usually the first thing we build with a new client, because it's fast to set up and the return is obvious within the first month.

2. The customers you've already written off

Every business has a database of old enquiries and past customers who went quiet — not because they weren't interested, just because nobody followed up before the moment passed. A progressive reactivation campaign works back through that list, re-qualifying old enquiries against today's pricing and availability, without anyone spending a week on the phone doing it manually.

Sydney Solar & Battery Company ran this against more than 1,000 dormant enquiries — some up to two years old — and recovered roughly $180,000 that the business had effectively already written off. An independent florist we work with generated $18,000 in tracked additional revenue in three months doing the same thing with occasion reminders and past-customer records. This is usually the highest-return automation for a business that's been trading for a few years, because the customers already exist — nobody has to go and find new ones.

3. The research you're doing after the enquiry lands

If someone on your team is manually looking up a new enquiry — checking the property, the company, past jobs, public information — before they can respond properly, that's time an automation can do in the background the moment the enquiry arrives. It doesn't replace the judgement call at the end; it just means the person making that call isn't starting from nothing.

For Simply Draft, this cut owner research time per enquiry from about 30 minutes to around five. That's not a marginal saving — it's the difference between responding to an enquiry same-day or three days later.

4. The hours nobody's meant to be working

After-hours calls, texts and emails are where the "should I answer this now or can it wait" decision costs business owners the most sleep, literally. The fix isn't answering everything immediately — it's triaging automatically, so a burst pipe reaches you straight away and a leaky tap gets booked in for the next business day, without you having to be the one making that call at 11pm.

This is genuinely running for a handful of our trade clients today, and it's less about chasing a headline number and more about a business owner not being the fallback for every after-hours message, urgent or not.

How to pick your first project

You don't need to solve all four at once, and you shouldn't try to. Pick the one project you can actually see and measure, and start there. A good first project usually has all of the following:

  • A visible cost. You can already point to it — missed calls, an ignored database, slow first response — rather than a vague sense that "we should be using AI more."
  • A clear trigger and action. Something specific happens (a call is missed, an enquiry lands, a lead goes quiet), and the response is well-defined enough to describe in a sentence.
  • Something you can measure. You'll know within weeks whether it worked — recovered leads, cut response time, extra bookings — not months.
  • No judgement call at the point of failure. If it goes wrong, a person notices quickly and picks it up — nothing quietly breaks and stays broken.

If a project you're considering doesn't tick most of those boxes, it's probably not the one to start with — not because it's a bad idea, but because you won't be able to tell if it worked.

What the first 30 days actually looks like

In practice, it's simpler than most owners expect. A 20-minute call establishes what's actually happening — where enquiries come from, what's currently being missed, and roughly what it's costing. From there, if there's a proven fit, the build itself is the fast part: connecting the automation to the phone number, CRM or booking system you already use, rather than asking you to run a second, separate tool.

Because these are repeatable builds, not first-time custom projects, most go live within a week of that first call. The next few weeks are just watching the numbers: missed calls recovered, leads reactivated, response time cut. If it isn't working, you'll know quickly — and that's the point of picking a first project you can actually measure.

What not to automate first

Leave anything involving real judgement, trust or a difficult-to-reverse decision until you've got a track record with the simpler stuff. That means holding off, at least for now, on fully automating things like pricing exceptions, complaint resolution, anything touching compliance or safety, or decisions about a specific customer relationship your team already understands better than any system will.

None of that is off the table forever — it's just not where the easy, obvious wins are, and a failed first project is the fastest way to convince a business that "automation doesn't work here" when the real problem was starting in the wrong place.

What this actually costs to start

One reason "just start with AI" advice is useless is that it never says what anything costs. Real numbers, for reference — these are the indicative setup-and-monthly rates for the four gaps above, not a quote:

  • Missed call text-back — from $750 to set up, $129/month to run.
  • Lead reactivation — $950 per campaign, one-off, not a subscription.
  • Lead intelligence (enquiry research) — from $2,200 to set up, $199/month.
  • After-hours triage — $950 to set up, $179/month.

None of these need a "digital transformation" budget or a multi-month project plan. Most are live within a week, because they're a proven, repeatable build rather than something invented from scratch for your business.

Where to actually start

If you're not sure which of the four gaps above is costing you the most, start with whichever one you can already picture happening this week: the call you know you missed, the enquiry that took three days to get a reply, the customer you meant to follow up with and didn't.

That's not a hunch. That's your first project.

Matt Wilson
Matt Wilson
Founder, momentuum

28 years building and running businesses across IT, digital and services — including 3 exits. Now helping other Australian business owners get their operations, automation and systems into shape. More about Matt →

Find the work in your business that should not exist

Before buying another AI tool, map the repetitive work happening across your business. Look at missed calls, slow lead follow-up, appointment administration, manual data entry and unanswered enquiries. A momentuum Automation Audit identifies where time and revenue are leaking, what can be automated safely and where human judgement still matters. No tech jargon. No obligation.

Start with what matters.

Tell us what feels slower, harder or less reliable than it should. We'll help you work out whether there's a worthwhile improvement — and the simplest way to make it happen.

No AI pitch. No giant workshop. No obligation.