What to Automate First (And Why Most People Get It Wrong)

Everyone asks the same question when they start thinking about automation: what should I automate? It sounds like the right place to start. It isn't. The better question is what in your business creates the most friction, takes the most time, and contributes the least to the work that actually makes you money. Start there. Everything else follows.

By Eva Petrakova · Local Business Systems

One thing first: none of this matters if nobody can find you yet. Automation speeds up a business that already has enquiries coming in. If you are not sure people searching for what you do can find you, here is how to check before worrying about what to automate.

The real test for what deserves to be automated

There are three things worth looking at when you're working out where to start.

First, where are you spending time that produces little return? Not every task that takes time is worth doing manually. Admin, follow-up, scheduling, answering the same questions, these eat hours without moving the needle.

Second, what do you consistently avoid or dread? Avoidance is information. If something keeps getting pushed down the to-do list, it's usually because it's tedious, repetitive, or both. That's exactly the profile of a task worth automating.

Third, what is pulling you away from the core work that earns you money? A personal trainer spends eight hours coaching sessions and two hours sending invoices, chasing replies, and rebooking cancelled sessions. The eight hours of coaching is the business. The two hours of admin is overhead. Automation belongs in the overhead.

The goal is to remove what distracts you, not to replace what you're good at.

The review request problem

A client hits a real milestone: the weight they had been working toward, the first pain-free week of a rehab programme, a personal best they had been chasing for months. They are genuinely thrilled, the kind of moment where they would happily tell a friend. Then nothing happens. Asking for a review in that moment feels awkward, especially when the milestone is something personal, and by the next day the moment has already passed.

That is exactly the profile of a task worth automating: repetitive, genuinely uncomfortable to do manually, low risk to the relationship with the client, and easy to fix without touching anything central to how you coach. I have written before about exactly how to set this up without it feeling awkward, so I won't repeat the mechanics here. The point for this post is just that it passes all three tests at once.

Content without the faff

Content is the other recurring example. Blog posts, Google Business Profile posts, a caption for a session photo: all of it takes time, and for most people running their own business, time is exactly what is already short.

The usual response is one of two mistakes. Either it gets skipped entirely, and a profile that goes quiet starts looking like a business that has stopped operating, which Google treats accordingly. Or it gets handed wholesale to an AI tool and published whatever comes out, which produces writing generic enough to describe any business in any town. Neither is automation done well.

What works is keeping the thinking yours and giving AI the specific inputs to work from, rather than handing it a blank brief. The opinions stay yours. The stories stay yours. What gets removed is you procrastinating because you don't know how to start.

The same logic scales down for shorter content too. A Google Business Profile post about what changed this week doesn't need a long voice note, it needs five minutes, not an evening you keep postponing.

This is the actual process I use for personal trainers and health coaches in Kingston upon Thames, if this is the layer you are stuck on.

The most common mistake: choosing the tool before knowing the need

Most people approach automation backwards. They see an ad for a platform, sign up, and then work out what to do with it. That's how you end up with an enterprise-grade system running a one-person business, and three hours of weekly admin just to maintain a tool that was supposed to save you time.

The tool is not the starting point. The problem is the starting point.

A powerful platform designed for a team of twenty carries assumptions about scale, workflow, and user capacity. Applied to a solo operation, it creates complexity rather than removing it. The admin doesn't go away. It shifts into platform management.

Choose the tool that fits what you've already decided to fix. Not the other way around.

The trap nobody mentions: automating what you actually enjoy

If you don't stop to think about which parts of your work you genuinely like, you risk automating the wrong things. You systematise the client conversations you looked forward to and stay stuck with the invoicing you've been avoiding for three years.

Automation should take what you dread and give you more time for what you're good at and what energises you. That's not just better for output. It's better for how you show up to the work that remains.

So before you automate anything, it's worth being honest about what you'd choose to hand off if you could. The answer shapes the priority list.

How to build the priority list

Personally, I don't start with a tool recommendation, I start with questions.

What does your business do, and what does delivery look like day to day? What's going well and what consistently goes wrong? If you could remove one thing from your plate tomorrow, what would it be? Where do you want to get better and currently can't? What do your clients say they need that could help the business grow?

That last question matters because there are two sides to this. One is efficiency: automating what you don't want to spend time on. The other is growth: doing things that weren't previously possible for you. Both are worth identifying before you decide where to start.

Then I map what you're currently doing, how you're doing it, and who's doing it. That mapping step is non-negotiable. You cannot sensibly prioritise automation without first knowing what you're working with. People skip it because it feels like admin before the real work. It is the real work. Get that wrong and everything built on top of it is guesswork.

If Google Business Profile is on your list and you want somewhere to start, here is the same checklist I use to set up every client profile from scratch, free, no obligation.

From there, the pattern is usually clear: start with what sits outside the core delivery of the service, where the volume is high and the task is repetitive, and where a mistake won't damage a client relationship. Get that running well. Then move to the next layer.

One tool that comes up in almost every one of those conversations is AI.

Is the average small business owner ready for AI?

AI is ready. It will keep evolving, but the best time to start using it is now. The real question is whether you know what to do with it.

You keep hearing: this tool is for this, that tool is for that. Then you find out all of them are paid. If you subscribed to every AI tool you see advertised for each separate task, you'd need to be earning double just to break even. Or you hear the advice to use AI as your assistant. Fine, it can write faster, and for some tasks that's all you want. But if you only use it through a chat window, you end up doing a lot of copy and paste, which is boring, and after a while you stop checking what you're pasting. At which point, who is the robot here?

So if you are staring at a chat window with a vague sense that you should be using AI better than this, that frustration makes sense. It means you've spotted the gap.

Part of the problem is prompts. You can write a good prompt if you already know what a good answer looks like. If you don't, AI gives you the most average response it can construct, and it will sound completely confident, including when it's wrong. It works by predicting which words follow which other words, based on what has been written. When the information it has is thin, it fills the gap with something plausible rather than something accurate. It will not tell you which is which.

Treat it as an assistant. And yes, I know, that's exactly what everyone keeps saying, usually without the how. So here is the how. Give it a specific task. Ask it to use only verified information, not inferred. Ask it to ask you questions until it's 95% confident it has what it needs to carry out the task. Ask it to be your sparring partner and challenge you, rather than agree with everything you say. Then check the result against what you actually know.

For SEO specifically, the trap is perfectionism. AI can help you move faster, but it can also pull you into overhauling everything at once (new content, new structure, new strategy) and you forget what you sat down to do in the first place. If the task is bigger, make a plan first: what steps, in what order. Follow the plan rather than every interesting distraction. Improving step by step is less exciting than disappearing down every rabbit hole. It also means you actually finish your projects.

The honest position is this: your experience, your knowledge of your clients, and your judgement about your work are still your competitive advantage. AI doesn't replace that. It multiplies the efficiency of everything around it, so more of your time goes to the work only you can do.

If you want help working out what this looks like for your business, Local Business Systems is the starting point.

© 2026 Local Business Systems. All rights reserved.