Training has never been faster to build. And it’s still missing the mark more often than it should. This is a practical guide to AI in learning design, where it genuinely helps, and where it quietly lets you down.
That’s the uncomfortable truth AI has surfaced. The bottleneck was never production speed. It was always the thinking behind what gets built.
This is a practical guide to where AI helps across the ADDIE model — Analyse, Design, Develop, Implement, Evaluate — and where it quietly lets you down.
The shift: AI makes it easier to build. Not easier to get right.
AI in learning design speeds up production at every stage of ADDIE, but it doesn’t replace the judgement needed to design learning that actually changes behaviour. There’s a growing belief that AI means you can do it all yourself.
Technically, yes.
Practically, not quite.
You can buy the tools. You can generate the content. You can follow the steps.
That doesn’t mean the outcome will land.
Quick example.
I can go to the supermarket and buy great steak. I’ve got a premium BBQ. All the right gear.
For a while, my daughter Coco loved my cooking.
Then I took her to Bistro Elba. Now my steak barely gets a look in.
Same ingredients. Same intent. Different outcome.
Why? Because the difference isn’t the tools. It’s the judgement, timing, experience, and feel.
AI in learning works the same way. Anyone can generate a module now. The hard part — knowing what good looks like, what learners need, and how to structure it so it sticks — that’s still human work.
We’re not saying, ‘ditch AI’. The question isn’t whether to use it. It’s knowing where it earns its place and where it doesn’t.
Get that wrong and you’re scaling mediocrity. Get it right and you’re producing better work, faster, with more time left over for the thinking that matters.
How we approach AI learning design at Hungry Minds
We’re not AI sceptics. We’re AI realists.
We use it across every stage of our work — but deliberately, not reflexively. There’s a difference between reaching for AI because it’s there and knowing exactly which part of the job it’s genuinely good at.
That distinction is what this guide is about.
We use ADDIE as our backbone — Analyse, Design, Develop, Implement, Evaluate — because it gives AI somewhere useful to sit. Each stage has different demands. Some are well-suited to AI assistance. Some need more human weight behind them. Knowing which is which is the skill.
ANALYSE
The Appetiser
Identify key learning priorities to deliver everything you need—and nothing you don’t.
No filler, no fluff.
DESIGN
The Recipe
Clarify scope, resources, and learning strategies with decision makers up front.
No surprises in your kitchen!
DEVELOP
Made To Order
We build engaging, interactive learning experiences— aligned to your goals.
Co-designed, iterated and perfected.
IMPLEMENT
Presentation
Prepare the client for a seamless roll-out in F2F, virtual, or blended environments.
Get pilot-ready and go for launch.
EVALUATE
Taste + Tweak
Review against learning outcomes —find out what to start, stop, and keep.
Identify changes for lasting impact.
Here’s how AI in learning design applies across each stage of the ADDIE model, and where human judgement still does the heavy lifting.
Stage 1: Analyse and Design
Where most projects win or lose
Analyse
Stop guessing. Start seeing.
AI is strong at pattern recognition. Feed it survey data, interview transcripts, existing training materials — it finds patterns fast. You can us it to:
- Analyse survey data in minutes
- Summarise stakeholder interviews
- Identify skill gaps across teams
- Audit existing training content
This speeds up discovery. It doesn’t replace it.
Speed isn’t the same as insight. AI doesn’t know what matters in your business. It doesn’t know which gaps are urgent, which stakeholders are resistant, or why the last training programme didn’t stick. That context is everything, and it only comes from actually talking to people.
Phase 1: Discovery and Direction
We run a focused discovery to understand your context.
This is where we do the work AI can’t. We get into the business — talking to stakeholders, understanding the commercial context, and asking the questions that don’t appear in a survey. What’s actually driving this need? What’s been tried before? What does success look like in six months, not just at sign-off?
AI can surface patterns in data. We figure out what those patterns mean and what to do about them.
What we do:
- Clarify your goals
- Review key materials
- Identify risks and gaps
What you get:
- A recommended approach
- Clear scope and timelines
- Defined investment — built on actual understanding, not AI assumptions
The thinking here is deliberate. Speed comes from clarity, not shortcuts (which end in rework).
Design
From blank page to structured thinking
AI is useful at the drafting stage. You can use it to:
- Draft learning objectives
- Map learning pathways
- Generate program outlines
- Suggest delivery approaches
But structure is not strategy.
This is where most DIY approaches fall over. They accept what AI suggests without pressure-testing it. Does this objective actually matter? Is this sequence logical for a real learner? Does this approach work for this audience, in this context, with this budget?
Phase 2: Design and Alignment
This is where the real value is created. We go deep.
What we do:
- Detailed learning needs analysis
- Stakeholder alignment
- Define learning outcomes
- Structure the full solution
- Develop a High Level Design
What you get:
- A clear, full learning blueprint
- Prioritised and structured content
- Defined learning approach and experience
High Level Design: Your blueprint for success
The HLD is not a draft. It’s a design asset.
It includes:
- Program structure and flow
- Learning activities and interactions
- Content breakdown and sequencing
- Delivery approach and formats
Why it matters:
- Aligns stakeholders early
- Reduces rework later
- Improves speed and quality
- Ensures the solution delivers outcomes
AI can help generate options. We decide which ones are worth building.
Stage 2: Develop, Implement, Evaluate
Where the work gets real
Develop
Speed creates volume. Volume is not quality.
This is where AI genuinely shines, and where the risk is highest. You can use it to:
- Generate scripts and modules
- Create quizzes and scenarios
- Repurpose content across formats
The trap is that fast output feels like progress. You end up with a lot of content that’s technically correct and experientially flat.
Another example. I bought a $900 spray gun to paint my fence. First few metres looked solid. Then it leaked. Paint everywhere. Three hours cleaning. Never finished the job. The following week I paid a painter to do the whole thing.
Same tools available to both of us. Different result.
AI can get you started. It can’t guarantee you finish well.
Phase 3: Development
We build in structured stages:
- Alpha: core build
- Beta: refinement
- Final: ready for delivery
You see progress at each stage. Feedback is focused and controlled. Someone with experience is making quality calls at every step.
Is this scenario realistic enough to create genuine reflection? Is this interaction adding to the learning or just breaking up the page? Does this module feel like something a real person would want to work through, or just something that covers the content?
Implement
From rollout to real-world use
AI can extend learning beyond the delivery event. You can use it to:
- Personalise learning pathways
- Provide chatbot support
- Automate nudges and reminders that bring content back into the workflow
Phase 4: Implementation
We support rollout with:
- Delivery preparation
- Facilitator support
- Technical setup
How a program lands matters as much as what’s in it. A rough implementation can undermine even well-designed content. A smooth one builds confidence in the learning before anyone’s clicked past the first screen.
This is also where we notice things. Early signals that something isn’t landing the way it should. Facilitators who need more support. Learner cohorts where drop-off is higher than expected. AI surfaces the data. We know what to do with it.
Evaluate
Stop measuring completion. Start measuring change.
Completion rates are not evidence of learning. They’re evidence of clicking.
AI removes a lot of the guesswork from evaluation. You can use it to:
- Track engagement patterns
- Identify drop-off points
- Surface insights quickly
- Predict performance gaps
The data is faster and richer than it’s ever been.
Phase 5: Evaluation and Improvement
Data tells you what happened. It doesn’t tell you why, or what to do about it.
We look at learner feedback, stakeholder insights, and performance data together — not as separate reports but as a connected picture.
We ask three simple questions:
- What do we stop?
- What do we start?
- What do we keep?
The answers aren’t always what the data suggests on the surface. Sometimes low completion points to a content problem. Sometimes a communication problem. Sometimes a culture problem that no course was ever going to fix. Telling the difference is experience, not analysis.
AI gives you the data. We turn it into decisions that actually improve the next iteration.
Where does AI go wrong in learning design?
Let’s be clear. AI is a multiplier. If your learning is weak, it exposes it faster.
The DIY approach looks efficient right up until it isn’t. Automated content that misses the point. Courses that are thorough but not useful. Learners who complete everything and retain nothing.
The issue isn’t the technology. It’s what you bring to it. AI produces better outputs when the thinking behind it is better. That’s true of every tool.
What this means for L&D teams
AI changes the job.
Less time creating content. More time solving problems.
Less building. More thinking.
If your value is production, AI replaces you.
If your value is judgement, AI amplifies you.
What good looks like
Analyse is sharper. Design is clearer. Development is faster. Implementation is embedded. Evaluation drives change.
AI doesn’t make learning better. Better thinking does.
AI helps you get there faster.
Explore more insights on AI learning design, instructional design, and training strategy:
https://hungryminds.com.au/blogs/