The 70% of Learning Was Never Designed For. AI Changes That.

Every L&D leader knows 70:20:10. But most organisations still design heavily for the 10%.

The model has been the backbone of talent development for decades. The idea is simple: 70% of learning happens through experience, 20% through others, and 10% through formal training.

And yet, when we look at where many organisations still place most of their learning investment, it is often the 10%. Workshops. E-learning modules. Offsites. Content libraries. Toolkits. These have value. But they are not where most capability is built.

The 70% — where people learn through experience, pressure, mistakes, feedback, and application — has often been left to chance. That is the gap AI can help close.

Why the 70% has always been hard to design for

It is not that L&D leaders do not understand the importance of experiential learning. They do.

The challenge is that experience is difficult to manufacture. You cannot create a high-stakes customer conversation just because a salesperson needs practice. You cannot engineer a difficult performance discussion simply to see how a manager handles it. You cannot let someone fail in front of a major client because they need learning reps.

So organisations did the next best thing. They trained people in a room, hoped the behaviour transferred, and measured attendance or satisfaction to show the program happened. It was never a true strategy. It was a workaround.

The bottleneck was never content volume. It was practice volume.

The conversation about AI in L&D is often about efficiency: faster content creation, smarter course recommendations, automated admin, and better learning search. These are useful. But they are not the most strategic use of AI in development.

Your people may know what good looks like. They may have attended the workshop. They may be able to recite the framework. But when a procurement lead says, "Your competitor is 18% cheaper — why should we stay?", knowing the framework and having the response ready are two very different things.

When a team member becomes defensive during feedback, a manager needs more than a model. They need composure, clarity, judgement, and practice.

AI changes this by creating a space where people can rehearse the hardest conversations they will face before those conversations happen in real life:

• Pricing pushback
• Performance feedback
• Stakeholder resistance
• Difficult team dynamics
• Coaching conversations
• Accountability conversations

All of these can now be practised repeatedly, safely, and with immediate feedback. That is the 70%. Finally designed for.

An inflection point in talent development

Organisations that continue to treat the 70% as something that "just happens" on the job will keep producing inconsistent capability. Some people will learn quickly. Others will avoid the difficult moments. Many will default to old habits when pressure rises.

Organisations that deliberately design for the 70%, using AI-powered practice as the bridge between training and live performance, can build capability at a speed and scale that was not previously practical.

The shift is not complicated, but it requires a different design logic. Not: workshop, hope, measure satisfaction. But: learn, rehearse, execute, reflect, repeat.

The early movers will not just train faster. They will build capability faster. The teams that practise more tend to perform better — not because they are automatically more talented, but because they have already made mistakes, received feedback, and improved before the stakes are real.

The 70% was always where development lived. AI is now making it designable.

The question for every L&D and HR leader is no longer whether AI belongs in development. It does. The better question is whether AI is being used tactically to create more content, or strategically to create more practice.