From prompt engineering to problem formulation

Welcome to the 48 new techies who have joined us since last Friday. If you haven’t already, subscribe and join our community in receiving weekly tech insights, updates, and interviews with industry experts straight to your inbox.

This week we’re quoting James Caan CBE (Serial entrepreneur, Media Commentator and Former Investor on BBC’s Dragons’ Den)

What Caan said:

“The problem is that we have this huge lack of skills; of people to be able to develop AI itself. I think the technology absolutely exists, but in order to execute and implement the technology and bring it into the workplace, there is a massive shortage of talent.”

We want to talk about prompt engineering

Prompt engineering in AI is, essentially, the process of teaching the AI what you want it to do. It learns as it goes – and your job is to help it learn what you want it to do.

Until recently, this would’ve been a pretty abstract concept for most people. But since ChatGPT exploded onto our desktops and smartphones, followed by a wave of other text-based generative AI tools, most of us have now become amateur prompt engineers – whether we know it or not.

Here’s an example

You open your AI tool of choice, and you write:

Tell me about LEAP.

The AI (we’re using Perplexity, in case you were wondering) responds:

A leap year is a year that contains an extra day added to keep the calendar year synchronised with the astronomical year or seasonal year.

But you didn’t want to know about leap years. You realise you haven’t given the AI enough information to provide you with a relevant response – so you need to make your request more specific.

So you write:

No, tell me about LEAP in Riyadh.

The AI replies:

LEAP is an annual tech convention held in Riyadh, Saudi Arabia, co-organised by Informa Tech, the Ministry of Communication and Information Technology (Saudi Arabia) (MCIT), and the Saudi Federation for Cybersecurity, Programming and Drones (SAFCSP).

That’s better. But actually, you’re looking for more persuasive information about LEAP. You want the AI to tell you whether or not it’s worth coming to the event. So, you ask:

What’s the best thing about LEAP?

And the AI gives you a list of bullet points detailing five of the benefits you can expect if you head to Riyadh in 2024, and closes with this:

Overall, LEAP in Riyadh offers a unique and valuable experience for startups, tech enthusiasts, and industry professionals. It provides a platform for innovation, networking, and knowledge sharing, contributing to the growth and development of the tech ecosystem in Saudi Arabia and beyond.

You type Great!

Perhaps a little confused by your praise, the AI throws another – shorter, punchier – version of bullet points about why LEAP is so good.

Then you go back to the beginning. You type:

Tell me about LEAP.

And what does the AI do?

It gives you six solid, conversational paragraphs about LEAP, including useful information and persuasive, benefit-driven text – and it even throws in some real photos from LEAP 2023.

What just happened?

Apart from taking the opportunity to remind you how great LEAP is, what just happened is prompt engineering in practice.

OK, so it was a bit haphazard. But by narrowing the focus of our requests, we taught the AI what we wanted – so that when we went right back to square one and asked the same question we started with, it gave us a completely different response.

Prompt engineering is a key AI skill (right now)

Right now, prompt engineering is a key skill for anyone who wants to leverage AI for their business. If you understand how to use prompts to nudge AI so that it gives you the information you need, you’ll be able to streamline work processes, so you can collect and use information more quickly.

But the reality is that AI will likely adapt fast, becoming more adept and more intuitive, and requiring less prompting skill from its users – or requiring different skills.

Like problem formulation: the ability to identify problems, analyse them, and describe them very precisely.

Writing for the Harvard Business Review, Oguz A. Acar (Professor of Marketing and Innovation at King’s College London) said:

“Problem formulation and prompt engineering differ in their focus, core tasks, and underlying abilities. Prompt engineering focuses on crafting the optimal textual input by selecting the appropriate words, phrases, sentence structures, and punctuation. In contrast, problem formulation emphasises defining the problem by delineating its focus, scope, and boundaries.”

In other words, prompt engineering is a linguistic skill. It’s about finding the right words.

And problem formulation is a comprehension skill. It’s about understanding the full scope of a problem, and then using language to communicate that problem accurately to the AI.

But to begin upskilling your team in AI…

Prompt engineering is a good place to start.

What AI skills are you working on right now? We’d love to know.

Open this newsletter on LinkedIn and tell us in the comments.

Have an idea for a topic you'd like us to cover? We're eager to hear it! Drop us a message and share your thoughts.

Catch you next week,
Richard McKeon
Marketing Director

P.S. - Mark your calendars for LEAP 2024 📅 4-7 March 2024. Want to be a part of the action?