About my approach
‘A jack of all trades is a master of none, but oftentimes better than a master of one.’
I’m Alejandra, an instructional designer who has been in this field long enough to see many trends come and go. Remember ‘learning styles’? Or when ‘badging’ was the height of gamification? I’ve lived through enough L&D revolutions to approach new ones with equal parts curiosity and scepticism. Which is exactly how I’m approaching AI.
I’m not going to tell you I’m an AI expert. Nobody with two years of experience in a field that barely existed three years ago should call themselves that. What I am is a good instructional designer who’s been paying close attention, experimenting carefully, and thinking hard about what AI can actually do and what it can’t.
The AI genie is out of the bottle. I’m not here to pretend otherwise, or to be an uncritical evangelist for it. I’m here to figure out how to use it thoughtfully, to stretch what’s possible in instructional design without losing sight of what actually makes learning work.
Why I don’t follow one framework
I’ve studied a lot of ID frameworks over the years. Action Mapping, Understanding by Design, ADDIE, SAM, Kirkpatrick, constructivist learning theory. I know them well enough to know that none of them do everything. Each one was built to solve a specific problem, and each one has blind spots.
The mistake I see a lot is practitioners picking one framework and applying it to everything, whether it fits or not. Good instructional design isn’t about loyalty to a methodology. It’s about understanding the problem in front of you and knowing which tools are most likely to help.
That’s the skill I’ve spent years developing. Not mastery of any one approach, but judgment about when to use which one.
The ID frameworks I draw on
Action Mapping
When the real need is behaviour change, not more content. Cuts through scope creep ruthlessly.
Understanding by Design (UbD)
When I need to design backwards from a meaningful outcome. Keeps the goal in view before a single slide is touched.
ADDIE
When a project needs structure and stakeholder checkpoints. Gives everyone a shared language for each phase.
SAM (Successive Approximation)
When speed matters and iteration is possible. Gets something real in front of learners fast.
Constructivist Learning Theory
When learners need to build understanding through experience. Informs how I design activities and reflection, not just content delivery.
Kirkpatrick Model
When I need to demonstrate real impact. Keeps evaluation designed in from the start, not added as an afterthought.
How I’m using AI, carefully
AI is another tool in the toolkit. A powerful one, and one that’s changing what’s possible in instructional design, but still a tool. I treat it the same way I treat ID frameworks: I try to understand what it’s good at, where it breaks down, and when to put it down and do something a different way.
In practice that means using techniques like chain-of-thought prompting to get more reliable outputs for complex ID tasks, self-consistency testing to sense-check assessments and learning objectives, and red teaming to find the failure points in chatbots and scenarios before learners do.
It also means being honest about what AI produces and what still needs a human. The judgment about what makes learning actually stick, that’s not something I’m outsourcing.
AI techniques I use
Chain-of-thought prompting
Getting AI to reason step by step through complex ID problems, rather than jumping to a surface-level answer.
Self-consistency testing
Running the same prompt multiple ways and comparing outputs to find the most reliable response, especially for objectives and assessments.
Red teaming
Deliberately trying to break or confuse an AI tool to find its failure points before learners encounter them.
Situational prompting
Matching the prompt structure and technique to the task at hand, not using the same approach for everything.
A note on how this page was written. The ideas here are mine, built up over years of practice and fairly strongly held opinions. The writing? That’s a collaboration. I used Claude to help turn a series of passionate rants about instructional design into something resembling coherent website copy. Which feels like a fitting illustration of the point: AI didn’t generate my thinking, but it did help me communicate it. That’s the kind of use I find genuinely useful.