AI

AI prompting for L&D: how to craft the perfect prompt

Omniplex Learning

Side shot of woman sat a desk, typing on a laptop

In this article [ hide ]

A new essential skill has emerged in L&D: knowing how to talk to AI.

“Prompt engineering” might sound like something reserved for developers and data scientists, but it’s now a core part of the toolkit for anyone using AI at work.

The principle is simple: better inputs mean better outputs. A well-crafted prompt can save hours, while a vague one can trigger so much back-and-forth that it would have just been quicker to do the task yourself.

If you often find yourself tapping “make this better” onto your keyboard, it’s time to sharpen your prompting skills. Here’s how to start.

 

What is prompt engineering in L&D?

Prompt engineering is the skill of knowing what to say to AI – and how to say it – to get the most useful response possible.

For L&D teams, getting the right prompt can lead to some seriously exciting – and time-saving – results. Imagine turning a dense policy document into a digestible lesson draft in minutes, or having a first-pass Vyond video script ready before your coffee goes cold. AI can speed up the process – but it all starts in one place: the prompt.

 

Why prompting skills matter for L&D teams

L&D teams are often under pressure to do more with less – faster turnarounds, leaner budgets, with no lowering of expectations. AI can help, but only if you know how to use it. That’s where prompting comes in.

There’s more to it than just saving time day-to-day, though. As AI becomes more embedded in how L&D teams work, prompting is slowly becoming a professional skill – not just a nice-to-have or a novelty. Think of it less as a trend to watch, and more as a skill worth honing sooner rather than later.

 

How to write a great prompt: the key components

The core of effective prompting – in L&D or anywhere else – is specificity. The more information you give AI to work with, the richer and more relevant the response. Here’s what to include every time, whatever tool you’re working with.

 

1. Give AI a job role

AI doesn’t know who it’s talking to unless you tell it. Assigning a role at the start of your prompt – “you are an experienced instructional designer” or “you are an L&D consultant writing for senior leaders” – immediately shapes the tone, depth and framing of the response. It’s a small addition that makes a big difference. Without it, you’re essentially asking a generalist to do a specialist’s job.

 

2. Tell it the ‘why’

Don’t just tell AI what to write – tell it why. Who is the learner, and what do they already know? Is this for a new starter totally fresh to the business, or a senior leader who needs the fluff cut? What’s the business problem this learning is trying to solve? The more specific the situation, the less generic the output – and the less time you’ll spend wrestling it into shape afterwards.

 

3. Be precise about the output

Tell AI what you need – not roughly, be exact. A bullet list, a script, a table, a summary – it won’t know unless you say so. In L&D, often the format can matter as much as the content. A scenario-based assessment looks nothing like a course outline, which looks nothing like a lesson summary. Specifying exactly what you need upfront saves you a round of reformatting before you can use it.

 

4. Set the parameters

Think of this as the guardrails. Word count, video length, tone, audience level – these are the details that stop AI drifting in the wrong direction.

There’s a notable difference between opening Vyond and typing in “make me a compliance video on manual handling” and “make me a 90-second video on manual handling for warehouse operatives, in a conversational tone with no jargon, using a scenario between a new starter and their manager.” The second one gives AI less room for interpretation.

 

5. Iterate with intention

Getting a great output rarely happens in one go – and that’s fine. The mistake most people make is typing “make this better” and hoping for the best. Instead, treat each follow-up prompt like a specific brief. “Shorten this to 100 words.” “Make the tone less formal.” “Replace the second example with something more relevant to a retail context.” The more precise your refinement, the faster you’ll get to your end goal.

 

6. Ground the output with source material

This is where prompting gets powerful for L&D teams. Rather than starting from scratch, give AI something to work with – an SME transcript, a brand tone of voice guide, an existing course or a policy document. This gives AI a thorough frame of reference – reducing the risk of it churning out inaccuracies, while keeping things on-brand.

 

AI prompting in action: L&D examples

The best way to see the difference good prompting makes is to see it in action. Here are five examples – from weak prompt to strong – across some common L&D use cases.

 

Compliance

Weak prompt: “Make a compliance video on fire safety in the office.”

Strong prompt: “You are an experienced instructional designer creating safety content for a large retail business. Write a 90-second microlearning video on fire evacuation procedures for office-based employees who have never received formal safety training. Use a scenario between a fire warden and a new starter. Plain English, calm tone – the goal is confidence, not fear.”

 

Onboarding

Weak prompt: “Write an onboarding module to welcome new starters.”

Strong prompt: “You are an L&D consultant designing onboarding content for a fast-paced retail business. Write a three-part onboarding lesson for new customer service representatives covering: what great customer service looks like in practice, how to handle a complaint, and where to go for support. Conversational tone, short paragraphs, no jargon, maximum 600 words.”

 

Leadership development

Weak prompt: “Create a knowledge check for a management leadership programme.”

Strong prompt: “You are an experienced leadership coach designing a reflective exercise for mid-level managers. Write five scenario-based questions, each based on a common people management challenge – such as handling underperformance or navigating a team conflict. Questions should prompt honest self-reflection, not right-or-wrong answers. Return as a numbered list.”

 

Sales enablement

Weak prompt: “Develop a sales training video on this new feature of a product.”

Strong prompt: “You are an L&D consultant supporting a B2B software sales team. Make a five-minute video introducing this new product feature to experienced sales reps who need to confidently pitch it to clients. Focus on the business benefit to the client, not the technical detail. Upbeat, confident tone.”

 

The prompt is just the beginning

One thing worth remembering: even a great prompt produces a first draft, not a finished product. AI can certainly move you through the mechanics faster, but it doesn’t get you the whole way there. The elements that really make learning stick still come down to human expertise, experience and judgement – no matter how good the prompt is.

 

Now you know how to prompt like a pro, it’s time to put that knowledge to work. Explore how Vyond can help your L&D team create better videos, faster. Get in touch to speak to one of our experts.

// -->