PROMPTING FOR LEARNING DESIGNERS: STOP ASKING AI TO ‘CREATE A COURSE’

When learning designers ask AI to ‘create a course’, they usually get something generic, bloated and disconnected from real learning needs. The problem isn’t the tool. It’s the prompt. This blog explains a simple way to structure prompts, so AI supports learning design work, rather than replacing it.

Imagine this familiar moment

You open an AI tool. You’re under time pressure. A stakeholder wants ideas fast.

So you type:‘Create an eLearning course on difficult conversations.’

The output arrives quickly. Objectives. Modules. Activities. Assessments.

At first glance, it looks helpful. Then you read it properly. It’s vague. It’s generic. It could apply to almost anyone, anywhere. And deep down, you know you couldn’t build a meaningful learning experience from it.

This is where a lot of learning designers quietly decide that AI ‘isn’t very good at learning design’.

In reality, AI is doing exactly what you asked. The issue is what you asked for.

Why ‘create a course’ is the wrong question

Learning designers don’t really design courses. They design decisions, behaviours and experiences. When you ask AI to create a course, you’re skipping the thinking that matters most. You’re asking it to jump straight to the output, without context, constraints or intent.

That’s why the results feel shallow.

AI is far more useful when it supports the thinking stages of learning design, not when it’s asked to replace them.

A more useful way to think about prompting

Instead of treating prompts as magic spells, treat them as design briefs. Good prompts usually contain four ingredients:

1. Context

2. Purpose

3. Constraints

4. The role you want AI to play

   
   

You don’t need to write more. You just need to be clearer.

Context: Ground the prompt in reality

AI doesn’t know your organisation, your learners or your business pressures unless you tell it.

Context might include:

·        Who the learners are.

·        What their role looks like day to day.

·        What’s changing around them.

·        What’s already in place.

This immediately moves the output away from generic learning theory and towards something usable.

Purpose: Be explicit about what you’re trying to solve

Learning design starts with a problem, not content. So instead of asking for a course, ask for help thinking through:

·        What’s the behaviour that needs to change?

·        What do people currently do instead?

·        Why does that gap exist?

AI is surprisingly good at helping you explore these questions, especially early in discovery or needs analysis.

Constraints: Tell it what not to do

One of the biggest mistakes in prompting is leaving the space too open.

Learning designers work with constraints all the time. Time. Attention. Tone. Risk. Compliance.

Including these in your prompt improves the output dramatically. For example, explaining that content must be short, scenario-led, suitable for busy managers or aligned to existing frameworks gives the AI something solid to design around.

Role: Don’t ask AI to be the designer

This is the most important shift. When you ask AI to ‘create a course’, you’re asking it to act as a learning designer. That’s rarely where it adds the most value. Instead, ask it to play a supporting role. A thinking partner. A challenger. A summariser. A pattern-spotter. A draft-starter.

This keeps the judgement, craft and accountability with the human designer, where it belongs.

What this looks like across the learning design process

During needs analysis

AI works best when asked to help surface possibilities, not answers. It can help you explore potential causes of a performance issue, suggest questions to test assumptions, or organise messy inputs from stakeholders. It struggles when asked to diagnose the problem on its own.

During SME interviews

AI is useful before and after the interview, not during. It can help you shape better questions, spot themes across transcripts, or reframe technical explanations into learner-friendly language. It shouldn’t replace the conversation itself.

During storyboarding

This is where many designers see the biggest time savings. AI can help you generate alternative structures, challenge linear thinking, or stress-test whether a narrative actually supports the behaviour change you want. But it still needs your judgement to decide what stays and what goes.

During assessment design

AI can suggest question types, identify common misconceptions, or help rephrase questions more clearly. It should never be the final authority on correctness, especially in regulated or high-risk topics.

Why this approach works better

When prompts are structured this way, something interesting happens. The output feels lighter. More tentative. More conversational.

That’s a good thing.

Instead of presenting itself as ‘the answer’, AI becomes a draft, a provocation or a starting point. Which is exactly how experienced learning designers already work with human colleagues.

A quiet shift in mindset

The most effective learning designers we see aren’t asking AI to do their job. They’re using it to:

·        Think faster.

·        See blind spots.

·        Get unstuck.

·        Explore options before committing.

They treat prompting as part of the design process, not a shortcut around it. And that’s why the results feel better.

FAQs

Why does AI produce such generic courses?

Because the prompt is generic. When context, purpose and constraints are missing, AI fills the gaps with the most common patterns it knows.

Do learning designers need to become prompt engineers?

No. They already have the skills. Prompting is just good briefing, something designers do every day.

Is it risky to use AI in early design stages?

It can be, if outputs are taken at face value. Used as a thinking aid rather than a decision-maker, it’s low risk and high value.

Should AI ever design a full course?

In practice, no. It can support parts of the process, but learning design still relies on human judgement, context and responsibility.

How do we help teams improve their prompting?

Focus less on ‘perfect prompts’ and more on shared principles: clarity, context and intent. The rest follows naturally.