How to Build AI Learning That Works for Everyone

When I worked at travel SaaS platform Deem, I noticed something that shifted how I think about user experience.

People would use our platform in entirely different ways — and not always how we intended. Some would breeze through workflows, others needed extra support, and a few found workarounds we hadn’t even thought of.

That’s when our team realized how critical it is to design for accessibility. Not just in how things look, but in how people think, learn, and interact.

Learning works the same way.

Especially when you’re rolling out something as big as AI. If you assume everyone learns the same way, you’re going to lose people.

That’s why AI training needs to account for neurodivergence — natural variations in how people process information and solve problems. This includes autism, ADHD, dyslexia and dyspraxia. According to Skillscast 15 to 20% of people are neurodivergent. That’s one in five on your team.

“The most interesting people you’ll find are ones that don’t fit into your average cardboard box. They’ll make what they need. They’ll make their own boxes.” — Dr. Temple Grandin

But most AI training still assumes there’s one “right” way to learn: slideshows, cookie-cutter modules, tech demos that expect everyone to just get it.

That setup fails a lot of people.

Not just neurodivergent learners but anyone who processes information differently.

When companies start recognizing neurodiverse strengths, the results speak for themselves. HBR research shows neurodivergent individuals can be more productive, are less likely to fall into cognitive bias traps, and often make decisions based on logic instead of gut feelings.

But they need the right setup to thrive, like structured environments, clear schedules, direct communication, and reminders that don’t feel patronizing. It’s not rocket science, just thoughtful design.

As a colleague of mine with dyslexia likes to say, “Don’t plan for me without me.”

And when it comes to learning AI tools and adapting to new workflows, the stakes are even higher. We’re not just asking people to learn what a tool does. We’re asking them to change how they work, and that’s a big ask.

The Right Way to Train Teams for AI (and Everything Else)

Most AI training programs fall short. They’re either too broad to be useful or too complex to put into action. In both cases, people walk away without changing how they work. When I build training, I focus on what gets results.

That’s why I rely on evidence-based learning.

This means giving people practical tools and real experiences that help them learn, and then allow them to apply that learning in the flow of their work.

Yale’s research backs this up. They’ve shown that people learn best when three things happen:

• Their prior knowledge and experiences are recognized and built on.

• They get time and space to practice in real-world settings.

• They have the opportunity to reflect and take control of their own learning.

What does that look like in practice?

Here are the methods I’ve found work best:

Active Learning

Instead of just feeding people information, active learning gets them involved — doing the work, having conversations, solving problems.

Some of my go-tos:

• Small group discussions where people share ideas and experiences.

• Problem-solving tied to real challenges they face at work.

• Collaborative projects that build both skills and team trust.

• Debates — because defending an idea forces deeper thinking.

This hands-on learning helps people build a solid understanding of the subject they can then use at work.

Formative Assessment

Training shouldn’t be a mystery.

Instead of waiting until the end to see if people “got it,” formative assessment is about checking in as they go.

Tools that work:

• Quick end-of-session reflections (“What’s clear? What’s still fuzzy?”)

• Think-Pair-Share activities to surface and address confusion.

• Questions that push people to explain or rethink their approach.

It’s low-pressure but helps people adjust and helps you spot gaps early.

Metacognition

When people start thinking about how they learn, everything shifts.

Metacognition gives them more control and helps them become independent problem-solvers.

Ways to support this:

• Asking them to predict outcomes and reflect afterward.

• Encouraging them to spot where they got stuck or what clicked.

• Tracking confidence levels during tasks.

Scaffolded Learning

People don’t master AI tools in one go. They need structure — a way to move from basics to complexity without getting overwhelmed.

Scaffolding helps:

• Break tasks into manageable steps.

• Offer examples before expecting independent work.

• Guide first, then gradually release responsibility.

• Use visual aids and templates for support.

• Leverage peer learning — sometimes help lands better when it comes from a teammate.

These strategies also support neurodivergent learners by offering more clarity, flexibility, and structure. And that’s the kind of learning environment where everyone performs better.

Designing AI Learning for Neurodivergent Minds

When I started building training programs, I used to think accessibility was mostly about visuals like alt text and color contrast.

But then I saw how many people struggled not because of how things looked, but because of how information was delivered.

That’s when I realized cognitive accessibility is just as important. If training isn’t designed with that in mind, it’s going to miss the mark and waste your team’s potential.

So, how do you design training that works for everyone, especially if they’re neurodivergent?

“For those of us who are neurodiverse, our learning differences are gifts that allow us to see the world differently and find solutions to complex problems. We are an asset to potential employers.” — David Flink

Here are a few easy elements you can incorporate into your programs:

Break up text

Big blocks of writing are hard to process. Short chunks, clear visuals, and white space make content easier for everyone. Skip narrated text slides — use graphics and voice instead.

Explain things more than one way

Some people want to hear it, others need to read it.

Lose the timers

Timed tasks stress people out and impact focus. If you have to use a timer, make sure people can pause it.

Tone down the noise

Flashing visuals, music, or high-contrast designs can be overwhelming. Give people control over those features or skip them altogether.

Be predictable

Clear agendas and instructions help reduce anxiety. If something changes, let people know early.

Let people work the way they want

Just because someone’s doodling or listening to music doesn’t mean they’re not engaged. Trust that they know how they learn best.

How I Build AI Training That Works

At one point, I watched a company spend months rolling out AI tools, only for them to sit untouched. A couple of webinars, a generic slide deck, and a PDF policy no one opened.

Leadership kept asking, “Why aren’t people using this?” The answer was simple: the training didn’t feel useful.

“Artificial intelligence is not a substitute for human intelligence; it is a tool to amplify human creativity and ingenuity.” — Fei-Fei Li

If you want to build a training program that people want to use, build in these elements:

Find your AI early adopters

Some of your employees are already experimenting with tools and workflows on their own. Ask who’s using what, and how? These folks are your best resource. Support them and learn from what they’re doing.

Build a small team

Bring together those early users, someone from IT, and L&D. This team can shape best practices, assess new tools and help define your AI policy.

Test small before scaling

Run a pilot with one department or team. Find out what works and what doesn’t, then roll it out to a wider audience.

Make the training interactive

No one wants to watch an hour-long presentation on AI. Create sessions that blend information sharing with discussion, idea sharing and hands-on practice.

Set up peer support

Pair up experienced users with beginners. Let them learn from each other instead of only relying on formal sessions. Studies also show that teaching someone else is a great way to improve your own knowledge and skills.

Keep it fresh

AI is advancing quickly. Update your materials often, offer live sessions, and send people to workshops or events so they’re aware of the latest strategies and tactics in your field. Keep the momentum going.

Track what matters

Forget course completions. Instead, assess behavior changes: are tasks getting done faster? Are outcomes improving? Are you saving on budget?

When training feels relevant and easy to apply, people use it. That’s when AI becomes part of everyone’s daily work.

I’m currently working with a client who invested a great deal of time and energy into their AI acceleration program. They assigned an owner, hosted informational workshops and ran a series of virtual calls where team members could pop in and troubleshoot their AI challenges.

We tracked the performance of the team over 12 months and are already seeing a $500,000 savings in budget and 2–4 hours per team member per week. But more importantly, the team is more satisfied with their daily work — now that AI is taking on the tedious tasks, they can do more of what they enjoy.

It’s a testament to a well-designed and thoughtful learning program.

AI Training That Sticks

Rolling out AI tools is the easy part. The real challenge is getting people to change the way they work.

“Change almost never fails because it’s too early. It almost always fails because it’s too late.” — Seth Godin

That shift doesn’t happen with a slide deck or a webinar. It happens when training is designed for real people with different strengths, experiences, and ways of thinking.

That’s why evidence-based learning matters. Using this framework helps people build new skills, apply them in real situations, and feel confident in their new work patterns. When training is active, practical, and flexible, it supports your whole team.

When you take the time to build AI learning programs like this you’ll start to see real change. Tools get used. Workflows improve. And people stop resisting AI and start owning it.