By Joanna Smith
There’s a meme going around LinkedIn that you may have seen. It compares the amount of time an athlete spends “training” vs “performing” and the amount of time a person in the workplace spends “training” vs “performing”.
As you can imagine, the two graphs are essentially reversed. The meme’s point is that we are expecting too much from people. That people need way more training before they can be expected to perform.
I think it misses the point. We humans are constantly learning, whether or not we’re in a formal “training” session or not. This is frequently expressed by the “70:20:10” concept popularised by Charles Jennings. The idea is that only 10% of workplace learning comes from formal training, 20% comes from social interaction with others, and 70% comes from figuring things out ourselves.
Micro-learning opportunities are everywhere. There are social observations like noticing when the boss is a little more stressed than usual. There are technical gems to be found like shortcuts in software. And there are procedural refinements thought of all the time. We are wired to constantly learn and improve our performance.
The challenge we face as L&D professionals is how to highlight, foster, and build resources for such continual learning. The most visible outputs of our profession have traditionally been courses. Whether in workshop format or eLearning modules, people expect us to produce them. How do we instead produce work that can harness the power of people’s natural curiosity and learning, without needing course completion statistics, or a specific set of assessment results to prove we did something?
If our clients (internal or external) have already decided that a course is their solution, it’s hard to suggest other types of intervention.
That had been my experience until a very recent project.
The intention was to develop some role-play practice activities that would sit inside a Storyline course. The client was a community support organisation, whose front-line workers went out and visited community members to provide support during difficult circumstances. The course was all about practicing the initial introduction. Certain elements need to come out during the first meeting, such as the role the support worker has in the situation, privacy principles and so on.
We designed four characters for learners to meet and have that introductory conversation with. Then we designed AI prompts to bring these characters to life, which would role-play with learners inside Storyline, and provide feedback. The simulated conversations were to be the final activity in the eLearning module, allowing learners to put into practice the skills we had presented earlier.
But what happened during testing surprised us.
During the testing process, we asked our client to work with a range of learners in the organisation. He did the activity with them, and asked them to share their thinking as they did it. We wanted to hear when they think something might sound ‘off’ or when they think the AI prompt is unrealistic, difficult, or simply uninteresting to work with. (Yes, we also invite them to ‘misbehave’ with the AI to ensure our guardrails are working).
With these particular learners, what we observed was quite different support work practice. None of it was necessarily ‘wrong’. There is no ‘right way’ to support a person. Much of a support worker’s practice depends on their experience, their tolerance for certain types of client reactions, and their preferences.
Even more interesting were the test simulations with experienced team members and managers. Discussions would ensue. Not with the AI simulation, but with each other. What does good practice actually look like? If someone felt they had performed a ‘textbook’ interaction, but were still provided with feedback on ways they could improve, does that mean their practice was no good? These peer-to-peer conversations were truly valuable.
Noticing that, our client made a significant decision.
He chose to separate out the simulation activity from the eLearning module, and instead, house it directly on the organisation’s SharePoint. He didn’t want the simulations to be hidden inside a course, where they would be relegated to that “10%”.
He wanted to give the activity space, make sure it was easily accessed, and he encouraged people to discuss it. Essentially, he promoted it to that “20%”.
This was an instinct our client had, and as it turns out, he was tapping into the same ideas that many other educators have had. According to Sellberg and Lindwall (2026), who looked at simulation-based training in professional education, “Across the studies, simulation emerges not merely as a tool for skill rehearsal but as a setting where professional knowledge itself is produced, negotiated, and reshaped...” (p.3)
Professional knowledge itself is produced through debate and negotiation. If this is the case, then producing an activity that stimulates debate and negotiation can be seen as a very important part of our job.
In our client’s context, it was impossible to ever have real practice observed. It goes against the principles of their practice to have a third-party present during client conversations. And while human-to-human role-play may be part of their initial training, it was a one-off, limited to that point in time in their professional learning journey.
This AI-powered activity was a welcome chance for practitioners to be observed, and get some feedback on potentially different ways to practice. It was clear from some that they wanted to go back and practice multiple times. Because the activity was powered by generative AI, when practitioners do it again, even with the same character, they get a unique experience every time. It seemed the activity was now being perceived as a “70%” kind of activity.
Our experience with this is not unique. Dr Philippa Hardman, when interpreting Synthesia’s 2026 AI in L&D Report, suggests that L&D teams worldwide are starting to realise the most powerful use of AI isn’t producing learning materials. It’s creating environments where learners actually practice. (Dr Phil’s Newsletter, March 2025)
Simulations are not new. And educators have always known (perhaps instinctively) that they work. Daneshfar and Karimi Moonaghi (2025) reviewed simulation-based training and concluded that simulation-based education can help strengthen clinical competence, decision-making, confidence, practical skills, empathy, and the integration of theoretical knowledge.
What is new, however, is the ability to provide simulation experiences in places learners can access them in their day-to-day work. They are no longer by nature relegated to the 10% environment of “formal training”. With AI-powered eLearning experiences, simulation can form part of the 20% and even the 70%.
References and Further Reading
-
Daneshfar, M. and Karimi Moonaghi, H. (2025) The impact of clinical simulation on bridging the theory–practice gap in nursing education: A systematic review, BMC Medical Education. https://link.springer.com/article/10.1186/s12909-025-07790-8
-
Hardman, Philippa (2026) March 25 edition of “Dr Phil’s Newsletter” https://drphilippahardman.substack.com/
-
Sellberg, C. and Lindwall, O. (2026) 'Simulation-based training in professional education: Learning, participation, and instructional design', Instructional Science, vol. 54, no. 9. https://link.springer.com/article/10.1007/s11251-025-09763-2
-
Synthesia (2026) AI in learning and development report 2026, https://www.synthesia.io/reports/ai-in-learning-and-development-report-2026
Interested in AI and eLearning?
AITD offers several courses on both AI and eLearning:
AI Essentials for L&D Professionals: Do you want to transform your L&D workflow and enhance learner experiences using Generative AI? In this blended learning course, you’ll learn how to use Generative (Gen) AI for a range of key L&D functions, including skills gap analysis, content creation, personalised learning and feedback. Register now.
eLearning: Foundations: This is Part I of an engaging, social learning suite of courses that provides you with access to learning experiences, activities and a comprehensive knowledge base. Register now. You may also be interested in eLearning: Planning and Design; and eLearning: Production and Delivery.
About the Author: Joanna Smith
Joanna Smith is the Managing Director of Pukeko Learning Solutions, a learning and development agency supporting organisations across New Zealand and Australia. Her work spans learning strategy, instructional design, eLearning, and facilitated learning. Her current focus includes using AI in thoughtful, fit-for-purpose ways. She is the immediate past President of NZATD.
