One of the reasons we brought the AI Practitioners Summit together was something I kept noticing: AI can seem so complicated that people hesitate to try it.
If that’s where you are, my advice is simple. Have a play. Pick one AI assistant, explain your business and something you’re struggling with, and ask where it could help. Then try one manageable suggestion.
You can start there, even if you have no idea what half the terminology means.
Because I think there’s something happening beneath all the talk about tools and technology. People look at everybody else and think, “They’ve got this all sorted. They’re clearly using AI to the nth degree, and I barely know where to start.”
That can bring a bit of embarrassment. Even shame.
You hesitate to ask a question because it feels like a question you should already know the answer to. And the longer you hesitate, the further behind you feel.
Meanwhile, the thing that might help you take a first step is sitting there, waiting for you to start a conversation.
Why does AI feel so overwhelming?
Some of AI is complex. Building reliable automations or connecting systems takes learning, testing and work. I’m not going to pretend otherwise.
But the complexity of those projects doesn’t tell you how difficult your first useful task needs to be.
There’s also incomplete information, misinformation and hypothetical advice to work through. You hear what AI might do, what it could change, and what someone thinks business will look like in five years. It can be harder to find someone showing you something useful you could try today.
And if you’re already comparing yourself with people who appear to be miles ahead, all that information can make starting feel even harder.
Skills are a real barrier for businesses. In an OECD survey conducted in late 2024 across more than 5,000 small and medium-sized businesses in seven countries, half of the businesses that weren’t using generative AI cited a lack of employee skills as a barrier. That finding comes from the OECD’s 2025 report on generative AI and the SME workforce.
So if you feel uncertain, there’s a good reason to learn. You can begin with something small enough to understand and check yourself.
What should you ask AI first?
I’d start by telling it about your business.
Choose an assistant such as ChatGPT, Gemini or Claude. Explain what you do, who you help and what you’re trying to achieve. Then describe a few of the bottlenecks in your day.
Ask it: “Where could you help me with this?”
Think about how you would brief someone who had just joined your business. You wouldn’t expect them to know what matters to you without explaining it. You’d give them context, talk through the work and let them ask questions.
That’s what I mean when I say treat AI like an employee. Give it a useful brief. Have a conversation. Tell it when it’s missed something.
You still decide what to use and what to do. The metaphor helps you communicate; it doesn’t mean the system has a person’s judgement or understanding.
Here’s a starting point you can copy:
I run a business that helps [type of customer] with [what you do]. I’m trying to [goal], but I keep getting stuck on [task or bottleneck]. I’m new to using AI. Ask me a few questions to understand the situation, then suggest some simple ways you could help. Explain what information you would need from me and whether each suggestion needs any extra tools or setup. Help me choose one small task to try first.
Use ordinary language. If an answer makes no sense, say so. If it suggests something too complicated, tell it you need an easier starting point.
You can also ask, “What have I left out that would help you give me a better answer?”
The suggestions are a starting point for discussion. Check that the assistant can actually do what it proposes with the information and tools available to it.
How do you choose your first useful task?
Look for something that takes effort repeatedly, and choose a task where you can easily judge whether the result is useful.
Maybe you spend too long drafting an email. Maybe you have an idea for a social media post but struggle to get it into words. Maybe you need help turning a rough list of notes into a clearer plan.
Start with a draft you can review before it goes anywhere.
The National AI Centre makes a similar point in its guide to choosing an AI solution: define what you want to improve, then choose the simplest option that meets your needs. It identifies standalone AI tools as an option for testing low-risk tasks.
Be clear about what the task involves. Drafting an email is different from sending one. Helping you interpret a supplied traffic report is different from accessing your website analytics. The second version may need permissions, connections or more setup.
For your first experiment, keep it manageable. Remove private details that aren’t needed, give it enough context to help, and read what comes back. Correct anything inaccurate. Ask it to change the wording if it doesn’t sound like you.
You may find that the conversation itself helps you get clearer about the problem.
The difficult email that showed me something more
I was already using AI heavily when this happened. It wasn’t my first discovery that it could help with a task.
I was having a particularly difficult ADHD day and realised I wasn’t in a state to record a podcast episode properly. I reached out to the person scheduled to be my guest, apologised and asked to reschedule.
The reply I received was incredibly rude. Borderline abusive. There was no empathy in it, and I’m not going to repeat the things they said.
I needed to respond, but I didn’t want to send an email filled with the emotion I was feeling.
So I explained the situation to my AI system. I told it what had happened, what the person had replied and what I wanted to say. I wanted a practical response that made it clear their behaviour had told me what I needed to know, and they would no longer be a guest on my show.
It helped me write that response. I sent it.
Then it did something I wasn’t expecting. It checked in with me. In essence, it acknowledged that being told not to take something personally wouldn’t necessarily stop me from doing so, and asked whether I was okay.
It almost made me cry.
I knew this was an AI system. It wasn’t a human being, and it didn’t have actual empathy. But the response landed at a moment when I was having a rough time, and it helped.
That was powerful for me. It broadened how I thought about the help AI could provide in my working day.
A difficult email can take up far more space in your head than the time required to write it. In that instance, AI helped me handle the communication clearly when I was struggling to do it myself.
Your first useful experience might be much more ordinary. It might be getting an awkward paragraph written or finding a sensible order for the things you need to do tomorrow. If it makes your day easier, that matters.
What should your first week with AI look like?
I’d like you to get to the end of that week with a little more time in your day, or a little less tension.
Being an entrepreneur is stressful. You have a lot to think about, and we’re constantly told we have to wear so many different hats. Some of those jobs can be done by employees or virtual assistants. Some can be helped by AI.
For the first week, give yourself a simple experiment:
- Describe one bottleneck. Choose something you genuinely want help with. Explain the situation and ask the assistant what it needs to know.
- Try a small task. Ask for something you can inspect, such as a draft or a plan.
- Give it feedback. Tell it what’s useful, what’s wrong and what’s missing. Keep talking until you have something you can use, or decide it isn’t helping with this task.
- Try the useful approach again. Save the instructions that worked and adapt them next time.
- Notice what changed. Did it reduce the effort? Help you get started? Save some time? Take a little pressure off?
You don’t have to master a collection of tools in seven days. Find out whether one of them can help with something that matters to you.
I’d love you to finish that week with a bit of relief and hope. A sense that you have another way to get some of the work done, and that you’re going to be okay.
Where can you learn from people actually doing this?
This is why we wanted to bring practical people together for the AI Practitioners Summit.
People who are really building stuff and really doing stuff with AI. People who can show you what they’ve tried, what they’ve learned and how they’re applying it in a business.
Once you start playing with AI, the world is your oyster. There’s so much you can explore. Having someone show you a useful starting point can give you that kind of leg up.
The summit runs 19–23 October 2026, with free registration and more than 60 practitioner-led sessions advertised. Choose sessions that connect to something you’d like to improve, and arrive with that problem in mind.
If you’ve been watching from the sidelines and wondering where to begin, register for the AI Practitioners Summit for free.
Bring your questions. That’s a perfectly good place to start.
