Why Prompt Engineering Feels Hard
(And What to Do About It)

I was halfway through a Gen AI training last week when I noticed something. We had just covered the basics of prompt engineering. Everyone nodded. It made sense to them. Give the AI context, be clear about what you want, and you will get better results. That sounded simple, until it was time to try it.

That's when I saw the pause. The hesitation. People were unsure about how to start. A few glanced at their colleagues around the room. It wasn't that they didn't understand the concept. They did. But putting it into practice felt strange, and a little unnatural.

It then occurred to me why. For years, we have worked a certain way. We search, we read, and we think. Then we piece things together ourselves. That is how we have always done things, and it has become second nature. Working with AI asks us to flip that around. Instead of doing all the research to build up our knowledge base first, we need to start with the end in mind. We need to describe what we want before we have done the work ourselves.

"It's like asking someone to delegate before they've proven they can do that task. It feels backwards."

And that's the real challenge, I think. Not learning the tools. Not even writing better prompts. It's unlearning the way we have been trained to work for so long.

In my trainings, we talk about Copilot, ChatGPT, Claude, prompt techniques, ethics, and governance. All of it matters. But what stays with me is the moment when someone lets go of the old way. Watching them trust the process, and seeing them leave with something they could actually use the very next day.

A Simple Shift That Helps

Here's something I've found helpful. Before writing your next prompt, pause first. Imagine you're handing this task to a colleague you trust. What would you tell them? What do they need to know? What does good look like?

That's your starting point, and the rest will follow.

The challenge with prompt engineering isn't technical. It's about shifting how we think about work itself. Once you see that, the whole thing becomes a lot less intimidating. You're not learning a new language. You're just learning to brief a very capable, very fast colleague.

What This Means for Your Team

If you're rolling out AI tools in your organisation, this is worth keeping in mind. The people who struggle most are not the ones who lack technical skills. They're often the most experienced professionals, the ones who have built strong, reliable ways of working over many years. Asking them to change that feels like a step backwards, even when it isn't.

Good AI training acknowledges this. It doesn't just teach tools. It gives people the space to recognise what they're unlearning, and to feel safe doing it. That's where the real shift happens.

The technology is ready. The question is whether we're giving people the right conditions to actually change how they work.