Your skepticism about AI content is correct.
Not the version that says AI will replace instructional designers. That’s a different argument and not the one we’re having here. I mean the specific, grounded skepticism that says: I have seen the outputs, I know what flat looks like, and I am not interested in putting something in front of learners that I’m not proud of.
That instinct is part of what makes you good at this job. The quality standard isn’t a barrier to using AI. It’s the prerequisite for using it well.
Here is what I am not hearing our industry say enough: AI in the hands of someone who knows what good looks like produces something very different from AI in the hands of someone who doesn’t. Your expertise isn’t in the way of this. It’s the whole point.
So, this post isn’t about convincing you that AI is worth it. It’s about showing you a path that doesn’t require you to lower your standard, commit to a philosophy, or spend time you don’t have on a learning curve that might not pay off.
Five places to start. All of them small. All of them low stakes. All of them useful.
1. Let AI draft your SME interview questions. Then edit them.
You know what you need to learn before you can design anything. You need to understand the performance gap, the context, the constraints, and what good looks like on the job. Getting there requires a good kickoff conversation with your subject matter expert.
Before your next kickoff, give AI the project overview, the stated request, and what you suspect the actual performance problem might be. Ask for a focused set of interview questions organized by what you need to find out.
Then edit them. Notice what it got right and what it missed about your specific context. The questions it gives you won’t be perfect, but they’ll be a starting point that saves you 20 minutes and usually surfaces one angle you hadn’t thought to ask about.
You’re still the designer. You’re still bringing the analysis. You’re just starting with more in your hand.
2. Ask AI what your learners might misunderstand. Use the answers.
Take the key concept from your current module. The one you’re about to build a scenario around. Describe it to AI and ask: ‘What are the most common ways someone misunderstand this concept or apply it incorrectly?’
Compare its list to yours. You’ll probably overlap on three or four. And there will likely be one or two that you hadn’t thought about addressing, not because you don’t know your content, but because you’ve been inside it so long that certain misconceptions have become invisible to you.
That’s the list your scenario needs to address. That’s where the branching decisions come from.
Your design instincts are what turn AI’s list into an actual scenario. But the raw input just got better.
3. Paste the compliance document. Ask for the behaviors.
You know this document; it is 5,267 pages long and uses regulatory language that no one understands. Legal or policy sends over to training and says, ‘we need a training.’ Your job is to figure out what employees need to do differently.
Paste the document. Ask: ‘What are the specific behaviors employees need to demonstrate to be compliant with this policy? What are they most likely to get wrong?’
Then compare AI’s answer to yours. It will pull behavioral language out of the policy faster than you can and give you a starting list of what the training needs to address.
You still must decide which of those behaviors warrants training versus a job aid versus a policy reminder. You still must decide what can be addressed in 20 minutes versus what requires deeper learning. Those decisions are yours.
But your starting list just got built in four minutes instead of forty.
Worth noting: This is also useful as a stakeholder conversation tool. When you can show a stakeholder the specific behaviors the compliance document requires and then explain which of those are training problems versus communication problems, you’re having a very different conversation than ‘I need three weeks to build this module.’
4. Have it write the course overview paragraph. Then make it yours.
This is a small one, but if you’ve ever stared at a blank document for twenty minutes trying to write the paragraph that goes at the top of a course before you’ve even built it, you’ll recognize the particular misery of this task.
Describe the course, the audience, the learning objectives, and the performance gap. Ask for a three-to-four sentence overview. You will need to rewrite it. It will be too generic. It won’t sound like you or your organization.
But it will give you something to push against, which turns out to be much easier than starting from nothing. Most people can edit in four minutes what would have taken twenty to build.
It’s not a trick. It’s just a better starting point.
5. Ask it to review your learning objectives. Actually listen to what it says.
Write your learning objectives. Then share them with AI, along with a description of the performance gap they’re supposed to address. Ask: ‘Do these objectives align with the actual performance gap? Are any too vague to be measurable? Is anything missing?’
This is the one that surprises people most. A well-prompted AI review will find at least one thing worth reconsidering. Not because your objectives are bad, but because writing objectives and reviewing objectives are different cognitive tasks, and having a second pass, even an AI second pass, catches things you’re too close to see.
You don’t have to accept every critique. You’re the one who knows the organizational context and the constraints you’re working with. But the review takes three minutes and it’s more useful than skipping it.
The line you’re holding.
You are not outsourcing your judgment. You are reducing the time you spend on tasks that don’t require it, so you have more of yourself left for the work that does.
The expertise that makes your work good: knowing your learners, diagnosing the real problem, pushing back on bad requests, editing for voice and quality, and making the call when AI gets it wrong, that expertise doesn’t go anywhere. It’s what makes any of this work.
Start with one of these five. Try it on your current project. Notice what changes and what still needs you.
That’s the whole experiment. Watch the VIP workshop this month in #IgniteLearning. We just did this live, it is recorded for you to watch, and shared fifteen prompts you can use as an in-house instructional designer.