I can talk about an SEO audit, improving search rankings or helping someone with AEO. But the person I’m trying to help might describe their problem much more simply: “I just want to get found on Google.”
That’s a different starting point for a conversation.
It’s also one of the opportunities I think business owners overlook with AI. We can use it to help investigate what people are asking, what they’re struggling with and the words they use to describe it. We can also look at the repetitive jobs inside our businesses that have become so familiar we barely notice them anymore.
Those are two places I’d look if you want to discover where AI could make a useful difference: what you need to understand about your customers, and what keeps taking time out of your day.
What could AI help you learn about your customers?
Research is one of the areas I’d really like more business owners to explore.
You might be researching a product, a service, an individual or a business opportunity. You might want to understand the challenges people face before you decide what to create for them.
An AI assistant with research tools can help find and organise relevant material. Accessible discussions on Reddit or Quora, search results, reviews and customer feedback can all give you clues about the questions people are trying to answer.
The assistant needs access to the material, though. Some sources won’t be accessible, and finding a handful of discussions doesn’t tell you what everyone thinks. Ask it to show what it found, where it found it and what it couldn’t check.
What interests me is the opportunity to get closer to the problem.
If people keep asking how to get found on Google, that gives you something to investigate. Why aren’t they being found? What have they already tried? What do they understand about the problem? What would a useful answer look like to them?
That can help with content ideas. It can also suggest a service, a resource or a different way to explain something you already offer.
You still need to test whether the idea is useful to the people you want to help. A discussion is a clue, not a promise that someone will buy.
Why do your customers’ exact words matter?
It’s easy to become fluent in the language of your own industry.
You understand the method, the terminology and the work involved. So you describe the service from that perspective. Your customer may be describing the result they want or the frustration they’re living with.
That’s the distinction in my SEO example. I might be thinking about an audit. They might be thinking about why nobody seems to find their business.
Both are relevant, but I need to understand their starting point if I’m going to communicate clearly.
One useful research task is to ask AI to collect the actual wording people use when discussing the problem you solve. Keep those words separate from the assistant’s interpretation of them.
For example, you could ask:
Research how [type of customer] describes [problem] in accessible public discussions and relevant search results. For each useful example, give me the exact wording, the source link, the date if available and enough context to understand it. Keep direct quotations separate from your summaries. Don’t invent quotations or claim you’ve accessed a source you couldn’t open. Identify recurring themes as possible findings to investigate, not proof of what all customers think.
If your assistant can’t browse, you can give it suitable material you’ve collected and ask it to organise that instead.
Then open the sources yourself. Check that the wording is there and that the context supports the interpretation. Someone complaining in a forum isn’t automatically a potential customer, and a frequently repeated phrase isn’t automatically a profitable opportunity.
You can take a possible theme into a real customer conversation: “Is this something you struggle with? How would you describe it? What would help?”
The useful outcome is a better understanding of the customer. That understanding can shape the next piece of content, the next question you ask or the next idea you test.
Which repetitive tasks have become invisible to you?
The other area I’d look at is the work you keep doing because you’ve always done it.
Some jobs become automatic to us. We copy information, organise things, answer familiar questions or prepare the same kind of material over and over. We don’t necessarily pause to ask whether the process could be easier.
If I were helping you recognise an opportunity, I’d ask:
- Where are the biggest time gaps in your day?
- What rabbit holes pull you away from what you’re supposed to be doing?
- Which recurring jobs drive you crazy?
Spend a few days noticing those moments. Write down the job and what makes it frustrating.
“Admin takes too long” is difficult to work with. “I keep taking the same information from an enquiry and organising it before I can reply” gives you a process to examine.
Or perhaps you repeatedly turn meeting notes into a list of actions. Maybe you prepare a regular report from information in several places. Those are examples to consider, rather than assumptions about what your business needs.
The National AI Centre’s guide to identifying opportunities recommends looking at business processes and pain points, then considering how well AI fits the task and the impact of an incorrect outcome.
That starts with understanding the work. Once you can see the steps, you can make a more useful decision about where help belongs.
How do you decide what to improve or automate?
Choose one recurring task and describe what happens from beginning to end.
What starts it? What information do you use? Which steps repeat? What result do you need? Where do you have to make a judgement or handle an exception?
You can ask AI to help you map that process before asking it to automate anything.
Here’s a recurring task in my business: [describe it]. Help me map its inputs, steps, output and exceptions. Ask questions where information is missing. Suggest which parts might benefit from a template, ordinary automation or AI assistance. Explain the setup and review each option would need. Don’t assume the whole task should be automated.
Some steps may be easier with a template. A straightforward rule might handle another. AI may help with interpreting text, organising material or preparing a draft that you review.
Look at the full effort involved. How often does the task happen? How much work would setting up the new process take? How much checking would it need? What happens when something goes wrong?
The National AI Centre’s guidance on measuring return on investment connects costs and benefits to business outcomes. That’s a useful way to think about whether an improvement is worth pursuing.
Try a small version with suitable example material. Compare it with your existing approach. Check the result and notice the extra effort involved in reviewing or correcting it.
If it helps, you have something to build on. If it creates more work, you’ve learned that before turning it into a bigger project.
Bring one opportunity to the AI Practitioners Summit
This is part of what excites me about bringing practitioners together.
Seeing what someone else is actually doing can help you recognise a possibility in your own business. You might arrive with a customer question you haven’t investigated properly, or a recurring job you’ve assumed you just have to keep doing.
At the AI Practitioners Summit, our headline speakers include Chase Aldridge, Sayre Blake, Punit Govind, Mark Kashef, Cole Medin, Hamish Prakash, Alexandra Spalato, Zubair Trabzada and Leon van Zyl, and I’ll be joining them too.
We’re bringing together people who are really building stuff and really doing stuff with AI. The event is focused on practical applications, with more than 60 practitioner-led sessions advertised across 19–23 October 2026.
I’d encourage you to bring one question from your business. Something you want to understand about your customers, or a task you’d like to make easier. Use that to help choose the sessions you explore.
There could be an opportunity sitting in the language your customers use, or in a job you’ve repeated so often you’ve stopped noticing it. Having someone show you a different approach can help you see it.
Register for the AI Practitioners Summit for free, and bring that question with you.



