AI is everywhere. Every new piece of software seems to have acquired an AI assistant, businesses are being told they need an AI strategy, and there is an endless supply of new tools promising to transform the way we work.
For a small business owner, that can make the whole subject feel considerably more complicated than it needs to be.
The useful question isn’t “How can I use AI?”
It’s “What am I doing in my business that could be done better?”
Start there, and AI becomes much more interesting.
Start with the problem, not the technology
One of the easiest mistakes to make with any new technology is finding an exciting tool and then looking for something to do with it.
I prefer to turn that around.
Think about the things in your business that take too long, get forgotten, cause frustration or require somebody to repeat the same task over and over again.
Perhaps enquiries arrive through several different channels and someone has to copy them into another system.
Maybe you repeatedly write very similar emails, prepare information in the same format, chase customers for documents or spend Friday afternoon compiling a report from information that already exists elsewhere.
Those are the places to start looking for opportunities.
And sometimes the answer won’t be AI at all. A simple automation, integration or change to a process may solve the problem perfectly well.
That’s fine too.
The objective is to improve the business, not to win a prize for using the most AI.
Can AI save you time?
This is probably the easiest opportunity for most small businesses to identify.
AI is particularly useful for repetitive tasks involving information that isn’t always presented in exactly the same way.
It might help to:
- summarise meeting notes and identify actions
- categorise incoming enquiries
- extract useful information from documents
- turn rough notes into structured information
- prepare the first draft of routine correspondence
- analyse feedback or survey responses
- create reports from existing business data
Individually, these tasks might save only a few minutes.
But a task that takes ten minutes and happens twenty times every week is suddenly more than three hours of somebody’s time.
That’s when small improvements become interesting.
Can you respond to customers more effectively?
Saving time doesn’t have to mean doing less for customers.
It can mean doing more, faster.
Imagine somebody completes an enquiry form on your website.
Instead of that information simply arriving in an inbox, it could potentially be categorised, added to your customer management system, assigned to the right person and acknowledged immediately.
A follow-up could be scheduled if nobody responds. Relevant information could be sent automatically. The customer doesn’t have to wonder whether their message disappeared into a black hole.
Not every part of that process requires AI. Much of it is straightforward automation.
AI becomes useful where some understanding or interpretation is required, for example identifying what an enquiry is about or summarising a lengthy message before passing it to the right person.
The best systems often combine AI, automation and human judgement rather than expecting one technology to do everything.
Can AI help with ideas and content?
Absolutely, but this is also where I think businesses need to be careful.
AI is excellent for brainstorming.
It can generate article ideas, explore different approaches to a subject, suggest questions customers might ask, organise information, or help turn a collection of thoughts into a sensible structure.
It can also help produce first drafts.
What it doesn’t possess is your experience of running your business.
It hasn’t met your customers. It doesn’t automatically know which of your opinions matter, what makes your service different or why something worked brilliantly for one client and disastrously for another.
Those are the things that make business content worth reading.
Use AI to help you communicate your expertise, not to manufacture expertise you don’t have.
Can it help you understand your business?
This is an area I think will become increasingly useful for smaller businesses.
Businesses accumulate information everywhere: website analytics, sales figures, enquiries, customer feedback, email marketing results, advertising data and spreadsheets.
Having the information isn’t the same as understanding it.
AI can help interrogate that data.
What are customers asking about most often? Which products tend to be bought together? Are there patterns in lost enquiries? Which subjects generate the most engagement? What changed between a good month and a poor one?
You still need to check the conclusions and understand the context, but AI can make exploring information considerably quicker.
Sometimes the most useful outcome isn’t an answer.
It’s spotting a question you hadn’t thought to ask.
What shouldn’t you hand over to AI?
Just because something can be automated doesn’t mean it should be.
I’d be particularly cautious about handing over decisions involving sensitive information, money, complaints, legal matters or anything where getting the answer wrong could seriously affect a customer.
There is also the question of data.
Before putting customer, employee or commercially sensitive information into an AI service, understand what information you are sharing, where it goes and how the provider uses it.
And always remember that AI can be confidently wrong.
A beautifully written answer isn’t necessarily an accurate one.
Human oversight matters.
Do you actually need another piece of software?
Possibly not.
Before subscribing to another AI service, look at the technology you’re already paying for.
Your email platform, website, CRM, accounting software, ecommerce system, office software and other business applications may already include useful AI or automation features.
You may also find that two systems you already use can be connected.
Adding another platform can sometimes create more work rather than less.
I would always rather make three existing systems work properly together than add a fourth because it has AI written on the box.
Where should a small business start with AI?
Find one thing.
Not an AI transformation programme. Not twenty new tools.
One irritating, repetitive or unnecessarily time-consuming part of the business.
Work out what currently happens, how much time it takes and what a better version of that process would look like.
Then ask whether technology can help.
That might involve AI. It might involve automation. It might require connecting two existing systems or changing the way information is collected in the first place.
Make the improvement. Measure whether it actually helped. Then find the next opportunity.
That’s how AI becomes useful to a small business.
Not by trying to become an AI business, but by using the technology where it makes an already good business work better.