Prompt Engineering Isn’t a Technical Skill. It’s a Thinking Skill.

Prompt Engineering isn’t a technical skill. It's a Thinking Skill.
By Sarah Williams

It wasn’t that long ago that I was standing in a room with a group of leaders in one of myprogrammes, confidently telling them that in the not-too-distant future they would be hiring prompt engineers, that this would become a defined role in their business, much like data scientists or business analysts have over the past decade, and that they’d need to start thinking about how and where those roles would sit across theirorganisation.

Boy, did I get that wrong!

Not because prompt engineering hasn’t become important, quite the opposite, but because instead of becoming a role, it’s quietly become a capability that sits underneath almost every knowledge-based role in the business, and increasingly, anyone who is interacting with AI in any meaningful way is now expected to be able to do this well, whether they’ve been formally trained or not.

What’s becoming really obvious, and I’m sure you’re seeing this as well, is that there is a growing gap between those who know how to prompt effectively and those who don’t, and that gap shows up everywhere. It shows up in the quality of outputs people are getting, the speed at which they’re able to work, the level of confidence they have when using these tools, and ultimately in the value they’re extracting from AI.

Some people are genuinely leveraging it to move faster and think better, while others are still getting fairly average outputs, or worse yet, using AI as a glorified search engine and wondering what all the hype is about.

And what I find particularly interesting is how familiar this all feels, because when I look at the core of what makes someone “good” at prompt engineering, it’s not actually new at all, it maps almost perfectly to the communication work I’ve been teaching leaders for years.

A big part of that work has been helping leaders ask better questions - because if you want better answers from your people, you need to be able to craft better questions in the first place! There is a depth of thinking required- you need to be clear on what you’re asking, why you’re asking it, and what kind of response you’re actually looking for.

And the same principle applies here.

At its core, prompt engineering isn’t really a technical skill; it’s a thinking skill, it’s about clarity, intention, and the ability to articulate what you want in a way that gives you something useful back.

Now, there are frameworks that can help, and Brynne Tillman’s CRISPY framework is a good example of a useful prompt structuring guide, particularly when you’re getting started or when you want to refine your approach.

But I don’t think the answer is that everyone now needs to go off and do a course on prompt engineering. In fact, one of the more interesting dynamics, as AI models continue to improve, is that you can quite literally ask the AI to help you improve your prompt. It's very meta - the system itself can play a role in developing your capability over time.

But even with that, the underlying skill still matters. Because if you don’t have a sense of what a good question looks like, you won’t recognise a good prompt when you see one, and you won’t know how to refine it to get closer to the outcome you’re actually after.

When businesses talk about developing prompt engineering capability, what I think they’re really talking about, whether they realise it or not, is developing a much more fundamental skill, which is the ability to think clearly and ask better questions.

And if we put our focus there, rather than getting overly caught up in tools and frameworks, what tends to happen is that the prompts naturally improve, the outputs improve with them, and the whole interaction with AI becomes far more valuable.

Which, when you zoom out, is exactly the same shift we’ve always been trying to create in leadership.