Instructional Design Skills That Will Be in Demand in 2026
The learning industry moves faster than we update our courses. Ironic, isn’t it?
Here are the skills that, in my view, already distinguish in-demand instructional designers from “slide makers.”
1. Prompt Engineering & AI Literacy for Learning
This isn’t about generating all content with AI. It’s about knowing where AI saves time and where human expertise is irreplaceable.
I’m not talking about “asking ChatGPT to write a quiz.” I’m talking about:
- Crafting effective prompts for realistic scenarios and case studies
- Validating AI-generated content for accuracy and learning objective alignment
- Integrating AI assistants into learning experiences themselves, not just the design process
No, AI won’t replace instructional designers. But instructional designers who work with AI will replace those who don’t.
The skill isn’t “using AI”. It’s knowing when to use it, when to override it, and when to put the laptop away and think.
2. Data Literacy: Learning Analytics That Actually Matter
Creating a course isn’t enough anymore. You need to understand what happens after people hit “enrol.”
Metrics that matter less than you think:
- Completion rates (people finish bad courses too, sometimes just to tick a box)
- Satisfaction scores (learners often rate “entertaining” higher than “effective”)
- Time spent (more time can mean engagement or confusion, the number alone tells you nothing)
Metrics that actually tell you something:
- Knowledge retention after 30/60/90 days, not just the post-course quiz, but whether it stuck
- Performance change: Are people actually doing the thing differently? Sales numbers, support ticket resolution times, error rates, whatever matters for this skill.
- Time-to-competency: how quickly can new hires perform independently after training?
- Assessment scores mapped to specific objectives, not just pass/fail, but which concepts people struggle with.
- Drop-off points: where exactly do people abandon the course? That’s your problem area.
What data-driven decisions actually look like:
- You notice 40% of learners replay the same video section three times. That content isn’t clear. Rewrite it or add a supporting resource.
- Assessment data shows people nail the concepts but fail the application questions. Add more practice scenarios.
- One version of your scenario-based module has 80% success rates, another has 50%. The content is the same but the framing is different. Now you know something about how your audience learns.
- Support tickets spike three weeks after onboarding training. Your course covers the basics but misses the edge cases people actually encounter.
Speaking the language of metrics is non-negotiable. Stakeholders want ROI proven with numbers, not hunches. “The feedback was positive” doesn’t cut it when budgets are tight.
3. Microlearning & Performance Support Design
Nobody wants 2-hour mandatory training anymore. People want information when they need it, where they need it.
Forget 40-slide courses. Think:
- Learning nuggets that take 2-5 minutes
- Mobile-first design (because that’s where people actually are)
- Role-based content personalisation
- Knowledge broken into bite-sized, actionable pieces embedded directly in workflow
We should no longer be asking “How do we cover everything?”, but “What’s the minimum someone needs to solve this problem right now?”
The best learning often doesn’t feel like learning at all. It feels like getting an answer exactly when you need it.
This works in practice. Walmart rolled out a microlearning programme for safety training across its distribution centres. Short, focused modules delivered through an app, with content that adjusted based on what each employee got wrong. The results: a 54% reduction in safety incidents, 91% voluntary participation, and a 15% increase in safety knowledge scores. Not completion rates. Actual behaviour change and fewer injuries.
4. Neurodiversity-Informed Design
15-20% of people are neurodivergent. That’s not just a drop in the ocean. That’s almost one in five.
Designing for different brains isn’t a nice-to-have anymore. It’s baseline:
- Multiple formats for the same content
- Flexible navigation (not everyone processes information linearly)
- Cognitive load management (less decoration, more clarity)
- Clear structure that doesn’t assume everyone thinks the same way
Design that works for neurodivergent learners works better for everyone. Clearer structure, less cognitive load, more flexibility. These aren’t accommodations, they’re good design.
Bonus: Facilitation & Blended Learning Orchestration
Here’s a paradox: the more digital learning we create, the more valuable live formats become. But not just lectures. I’m talking about facilitation, peer learning, practical workshops.
The instructional designers who stand out create blended experiences where online and offline work together, not in parallel:
- Async preparation that leads to meaningful sync collaboration
- Live sessions that couldn’t happen any other way
- Follow-up that reinforces and extends what happened in the room
“Watch these videos, then we’ll have a webinar” isn’t blended learning. It’s two separate things pretending to be connected.
Real blended design means each modality does what it does best. Pre-work builds shared understanding. Live time is for the things you can only do together. Async follow-up makes it stick.
This is more of a nice-to-have than the four skills above. Many ID roles are primarily digital, and facilitation is a different skill set entirely. Worth developing if your work calls for it, but not essential in the way AI literacy or data skills are.
What This Means
None of these skills are optional extras anymore. They’re the difference between IDs who shape how organisations learn and those who get handed a PowerPoint and told to “make it interactive.”
The job isn’t about building courses. It’s about solving learning problems, with whatever approach actually works.
