You’ve read the articles. You’ve heard colleagues talk about it. Maybe you’ve even played around with a chatbot at home, asking it to write a silly poem or settle a debate. But when it comes to actually using AI for something real at work, you haven’t quite made the leap.
If that sounds familiar, you’re in the majority.
Plenty of people are curious about AI but stuck at the “I wouldn’t know where to start” stage. Last month’s piece on the AI already woven into your daily life showed how much of it works quietly in the background. This month, we’re taking the next step: picking a real task and doing it.
Start Smaller Than You Think
The biggest mistake is choosing something too ambitious. You don’t need to automate a process or build a system. You need one small, low-stakes task you’d normally do yourself – something where a mediocre result doesn’t matter because you’re going to check the output anyway.
Good first projects: drafting a reply to a routine email, summarising a long document, brainstorming ideas for a presentation, tidying up meeting notes, or getting a quick explanation of something unfamiliar. Notice what these have in common – AI does the heavy lifting on a first draft, and you refine it.
How to Actually Do It
Pick your task, then open a free AI chatbot – ChatGPT and Claude are both good starting points. Type what you need in plain English. No special commands required.
For example:
“I need to write a short email to a client explaining that their delivery will be delayed by a week. Keep it professional and apologetic.” Or:
“Summarise the key points from this text in five bullet points”
– then paste in the content.
The AI will produce something. It probably won’t be perfect first time, and that’s fine. The skill is in refining your request. If the tone is wrong, say so: “Make it less formal.” If it’s missed something, add it. Think of it as directing a keen but slightly literal colleague.
One thing worth knowing while you experiment: anything you type into a free AI tool could be stored or used to train the system. So avoid pasting in confidential company information, personal data, or anything sensitive. Stick to general content – there’s plenty to try without going near anything private.
Try This: Set aside 20 minutes this week for your first experiment. Pick one task from the list above, try it with a chatbot, and note honestly – did it save you time? Was the output useful? What would you do differently next time?
Checking the Output (The Important Bit)
AI tools are confident – they’ll give you a polished response even when they’re wrong. Always read what comes back with a critical eye. Check facts. Adjust the tone to match your voice. Remove anything that doesn’t sound like you.
The goal isn’t to hand over your thinking. It’s to speed up the routine parts of your work so you have more time for the parts that need your actual expertise and judgement.
Pro Tip: Tell the AI what role to play and you’ll often get better results. Try: “You’re helping me draft professional communications for a UK workplace” or “Act as a plain-English editor.” It focuses the response surprisingly well.
Where the Real Value Lives
People who stick with AI tools find the biggest gains in drafting (emails, reports, summaries), research (quickly getting up to speed on unfamiliar topics), and brainstorming (generating options you can then evaluate). None replaces your expertise – they shrink the gap between a blank page and a working draft.
You don’t need permission to start. You don’t need a company policy or a training course. You need 20 minutes and a willingness to experiment. The worst that happens is you delete what it produces and carry on as before – but with a clearer sense of what’s possible.
