Are you looking at automation for your business but struggling to work out what’s actually worth automating and what isn’t?
You’re not alone. Automation has become one of those words that gets thrown around so often it’s started to lose meaning. Everybody’s telling small business owners to “automate everything,” but nobody’s explaining how to actually decide what’s worth the effort.
We get asked about this constantly. Someone’s read an article, watched a video, or had a chat with a mate who “automated their whole business” (their words, not ours), and now they want to know where to start.
The honest answer is that most small businesses are automating the wrong things, for the wrong reasons, in the wrong order.
This article is about fixing that. Not with theory. With a practical way of working out what genuinely deserves your automation budget, and what should stay firmly in human hands, at least for now.
Before We Go Any Further
This is simply our view, based on what we’ve seen working with small businesses over the years.
We’ve watched businesses spend serious money automating something that barely mattered, while the actual time-drain in their business carried on exactly as before.
Have a look around, speak to other business owners, and make your own judgement. This is just how we see it.
What Actually Makes a Task Worth Automating?
A task is worth automating when it’s high-frequency, low-judgement and rule-based. That’s the whole test.
High-frequency means it happens often enough that the time saved actually adds up. Low-judgement means there’s no real decision involved, just a repeated action. Rule-based means you could write the steps down and someone with no context could follow them exactly.
If a task hits all three, it’s a strong automation candidate. If it’s missing even one, you need to think a bit harder before touching it.
The trap most businesses fall into is automating something because it’s annoying, not because it’s actually worth automating. Annoying and worth automating are not the same thing.
What Does “The Five-Minute Task You Do 100 Times a Month” Actually Look Like?
This is the sweet spot. A task that takes about five minutes but happens roughly a hundred times a month is quietly costing you over eight hours. That’s a full working day, every single month, spent on something with no real decision-making in it at all.
In our experience, these tasks tend to look something like:
– Chasing overdue invoices with the same reminder email
– Copying a new enquiry from a web form into a CRM or spreadsheet
– Sending appointment confirmations and reminders
– Generating the same weekly or monthly report from the same data source
– Updating a spreadsheet every time a specific event happens, such as a sale, a booking or a stock movement
– Forwarding certain types of email to the right person or folder
– Creating a standard onboarding checklist every time a new client or employee starts
None of these need a human to sit and think about them. They need a human to type the same thing, in the same order, over and over again. That’s not really a job. That’s a job someone gave a computer by accident, years ago, and nobody’s got round to fixing since.
“I’m all for automation. I’m less keen on spending £10,000 automating something that took somebody five minutes in the first place. Start with the boring, repetitive stuff. That’s where the real time comes back.”
What Shouldn’t You Automate Yet?
Anything that genuinely needs a human to weigh things up shouldn’t be automated, at least not fully.
That includes things like:
– Deciding how to respond to a genuinely upset customer
– Pricing a bespoke quote where every job is slightly different
– Making a judgement call on whether to chase a client who’s a bit late paying, versus one who’s clearly in trouble
– Interviewing candidates or making hiring decisions
– Handling a complaint that involves reputation, not just process
– Any conversation where tone, relationship or context genuinely changes the right answer
The reason is simple. These tasks don’t follow a fixed rule. They follow “it depends,” and “it depends” is exactly the kind of thing that goes badly wrong when you hand it to a system that can’t actually think.
To be fair, some of these can be partly automated. You might automate the alert that a complaint has come in, without automating the actual response. That’s a different thing entirely, and we’ll come back to it shortly.
A Hypothetical Example, Just to Illustrate the Point
We don’t have a specific client case study to share here, so let’s use a hypothetical example instead, because it’s a pattern we genuinely see repeated across different types of business.
Imagine a small accountancy practice decides to automate its client communication entirely, including responses to queries about individual tax situations. The system sends confident, professional-sounding replies. The problem is that some of those queries need actual judgement, because every client’s situation is slightly different, and a generic answer applied to a specific tax question can be actively wrong rather than just unhelpful.
The lesson isn’t “don’t automate anything in a professional services business.” It’s that automating the wrong layer of a task, the bit that actually needs a brain, causes more problems than it solves.
What About the Tasks Sitting in the Middle?
This is where most of the genuinely useful automation actually happens, and where most businesses give up too early.
A lot of tasks aren’t purely rule-based or purely judgement-based. They’re a mix. The trick is separating the two halves.
Take that complaint-handling example again. You probably shouldn’t automate the response. But you could easily automate:
– Flagging that a complaint has come in
– Logging it against the right customer record
– Notifying the right person immediately
– Tracking how long it’s been open
– Reminding someone if it hasn’t been actioned within a set time
None of that requires judgement. All of it currently relies on someone remembering to do it manually, which is exactly the kind of thing that gets missed on a busy day.
This is the real skill in automation. Not automating everything. Splitting a messy task into its mechanical parts and its judgement parts, then only automating the mechanical half.
“Sometimes the cleverest piece of technology you can introduce is the one that simply saves somebody an hour every day. It doesn’t need to be clever. It just needs to actually work.”
So, What Should You Actually Do?
Here’s the practical version.
- Write down every repetitive task in the business for a week. Genuinely write it down, don’t rely on memory.
- Estimate how often each one happens and how long it takes.
- Ask whether it involves any real judgement, or whether it’s the same steps every time.
- Multiply the time by the frequency. Anything eating more than a few hours a month with no real judgement involved goes to the top of the list.
- For tasks that are a mix of mechanical and judgement work, split them. Automate the mechanical half only.
- Start with the single biggest, simplest win, not the most impressive-sounding one.
- Test it properly before rolling it out everywhere.
- Measure whether it actually saved the time you expected. If it didn’t, work out why before automating anything else.
Checklist: Is This Task Worth Automating?
Before automating anything, ask:
– Does this happen often enough that the time saved genuinely matters?
– Could someone with no context follow the exact same steps every time?
– Is there a real decision being made, or just a repeated action?
– What happens if the automation gets it wrong? Is that a minor inconvenience or a serious problem?
– Can a human still check or override it easily?
– Have we actually written down what “correct” looks like for this task?
– Would we be embarrassed to explain, in plain English, what this automation actually does?
If you can’t answer these clearly, you’re probably not ready to automate that particular task yet, and that’s absolutely fine.
Frequently Asked Questions
What is the best type of task to automate first in a small business?
The best place to start is a task that happens frequently, follows the same steps every time, and involves no real judgement. Chasing invoices, sending standard confirmations, and moving data between systems are typical examples.
How do I know if a task is too complicated to automate?
If the honest answer to “what should happen here” is regularly “it depends,” the task probably involves judgement that shouldn’t be fully automated yet. You can often still automate the surrounding admin, just not the decision itself.
Can I automate customer service without losing the personal touch?
Parts of it, yes. Automating confirmations, reminders and routine questions tends to work well. Automating the actual handling of upset or complex customers tends to backfire, because it removes exactly the judgement that customer relies on.
How much time should a task be taking before it’s worth automating?
There’s no fixed rule, but a task consuming several hours a month with no real judgement involved is usually worth investigating. A five-minute task done a hundred times a month adds up to over eight hours, which is where the maths starts making sense.
Is it worth automating something that only happens occasionally?
Usually not, or at least not first. Rare tasks rarely justify the setup time and cost. Focus your automation effort on things that happen often enough to actually move the needle.
Do I need expensive software to automate small business tasks?
Not always. Plenty of the highest-value automation in a small business uses tools you may already have, connected together sensibly. The cost should relate to the value the automation actually delivers, not the other way round.
Our Final View
Automation genuinely can save small businesses a serious amount of time. We’re not against it. We use it ourselves.
But the businesses getting the most out of it aren’t the ones automating the most things. They’re the ones automating the right things, in the right order, starting with the boring, repetitive, low-judgement work that’s been quietly eating hours out of every single week.
If a task needs a human to think, leave it with a human. If it just needs a human to repeat the same steps, give it to a machine and get your time back.
Wondering What You Could Actually Automate?
Spending hours doing something manually and wondering whether it could be automated? Give us a call and talk us through the process. We’ll tell you honestly whether it’s worth automating, or whether your time’s better spent elsewhere.