The Designer You Did Not Know You Needed

By Michael Ajala

In the last two years, almost everyone I know has tried to teach themselves something new. A friend in Lagos picked up Python after losing a job to a graduate who knew it. A cousin in Abuja moved from accounting into data analytics through a six-month bootcamp. Three designers I mentor have started using AI tools so heavily that their workflow has fundamentally changed. The upskilling wave is real, and with the momentum it has built, it shows no sign of slowing down.

And yet, when I look at the products built to support all these learning bootcamps, online courses, in-app tutorials, and the millions of YouTube videos—that have, in many ways, effectively become Nigeria’s adult education system—the design quality is, mostly, poor. The information is there. The structure is not. Users are forced to assemble their own curriculum from scattered resources and most often get frustrated and quit before they finish.

This is the gap that instructional design fills. And it is steadily becoming one of the most important roles on the product teams I work with.

Most people in tech first meet the term “instructional design” in the context of corporate e-learning, the slightly dated discipline behind compliance modules and onboarding videos. That association is unhelpful. Instructional design today is a much broader practice: the craft of taking complex information and making it easy to learn, in context, at the moment a user needs it. It overlaps with UX writing, onboarding design, content strategy, and increasingly, conversational AI design. It is the layer that decides whether your user learns to use your product, or gives up two screens in.

I work on this directly in my current role. Part of my job involves designing experiences that help senior leaders at a large bank find and engage with talent and learning support. The user base is busy, senior, and impatient. If a feature does not teach itself in the first 15 seconds, it loses them forever. Adoption rose ten per cent in the second half of last year, not because we added more features, but because we made the existing ones legible. That is instructional design wearing a different name.

At a US digital healthcare platform I previously led design, we increased onboarding completion by 35% through a similar approach. We stopped trying to explain the whole product upfront. We taught one task at a time, in the moment the user actually needed it, with as few words as we could get away with. Most of the improvement came from removing instructional content, not adding more of it. The best instructional design is invisible.

What makes this discipline urgent now is a combination of three forces, all converging at the same time. The first is the explosion of new tools. AI alone has introduced an entirely new way of working. The products being built around it, including the dozens of GPT wrappers African startups are quietly shipping, all assume users already know how to prompt, evaluate model outputs, and integrate AI into their workflows.  None of those assumptions holds for most users. The products that succeed will be the ones that teach as they go.

The second is the collapse of the traditional learning pipeline. In Nigeria, formal education has not kept pace with the skills the market actually demands. People are learning outside institutions: through bootcamps, YouTube, paid courses, and increasingly, through the in-app tutorials of the products themselves. The product is now also the school. A designer who cannot think about how the user learns is shipping half a product.

The third is the rise of AI-assisted upskilling. AI tools like ChatGPT, Claude, and a dozen other tools have made it possible for an individual to learn a new technical skill faster than ever before. But the quality of that learning depends entirely on the structure of the interaction. Without an instructional design layer, AI tutoring degenerates into the same scattered, unstructured information overload that drove people to seek out tutoring in the first place. The model knows the answer. It does not know how to teach.

Companies are slowly starting to hire specifically for this role. Major tech companies have started building out “learning experience” teams. Healthcare and fintech products, in particular, have started to recognise that user onboarding is not a single-sprint task but an ongoing design discipline. In my observation, African product teams are behind on this. Most still treat onboarding as a checklist item handed to a junior designer two weeks before launch, with the brief “make sure new users know what to do.” That is not instructional design. That is a tutorial, and it is a different thing.

There is a competitive argument for African startups to take this seriously. The Nigerian fintech market is now largely saturated. Feature parity is the norm. What separates the products that survive is whether their users can actually use them, especially first-time users who are not particularly technical and have limited patience for an app that does not explain itself. A bank app that teaches its users how to set up a USSD shortcut in three taps is going to retain that user better than one that ships an identical feature buried in a settings menu. In many cases, the teaching becomes part of the product itself.

There is also a personal argument for designers reading this. If you are a product designer trying to figure out what to specialise in over the next five years, learning design is one of the most underserved and high-impact directions you can take. It does not require new tools. It requires the discipline to think rigorously about how a stranger encounters your product and what they need to understand, in what order, to succeed. That is design work in its purest form.

The era of upskilling has produced an enormous number of new learners, all looking for products that respect their time and meet them where they are. The products that win their attention will not be the ones with the most features. They will be the ones who teach the most clearly. Someone has to design that. It might as well be you.

Also read: Go Systems design at scale: Rethinking software engineering in the AI era