Showcasing Learning Design Strategies: Using AI Avatars to Support High Stakes Role Based Learning

Every fortnight, our team runs an internal showcase to explore how learning design decisions show up in practice. This edition focused on the use of AI avatars to create scenario-based learning videos, and how simulated dialogue can help learners prepare for important, structured interactions where confidence and clarity matter.

Rather than using video to explain content, the showcase explored how AI-generated role play can help learners understand what to expect, how conversations flow, and what good preparation looks like before they step into a real situation.

The Focus: Preparing Learners for Structured, High-Stakes Interactions

This showcase focused on the design of short learning videos created using AI avatars, embedded within an eLearning experience. The videos centre on a formal interaction that learners are expected to participate in as part of their role.

The scenario involves two clearly defined roles. One represents an external assessor or reviewer. The other represents the person responsible for presenting, responding, and engaging on behalf of their organisation. The interaction follows a defined structure and flow, with expectations on both sides.

The intent of the learning was not to teach content, but to build readiness. As discussed during the showcase, seeing the interaction unfold. 

Learners can observe how the conversation progresses, what kinds of questions are asked, and how responses are framed.

“It’s almost like showing them a bit of a simulation of that experience.” 

This simulated exposure helps learners feel more prepared and less reactive when they encounter the real interaction.

The Design Challenge: Making Role Play Scalable and Credible

Role play is a powerful learning method, especially for situations that are formal, structured, and potentially stressful. However, it is also resource-intensive and difficult to scale.

The design challenge was how to create role-based learning that feels credible and human, without relying on live role play or filmed actors. Learners needed to see a realistic exchange, not a checklist or script, but the solution also needed to be flexible enough to evolve.

This led to the use of AI avatars as a way to model dialogue between roles, rather than as presenters delivering information. The focus was on interaction, tone, and flow.

Design Strategy 1: Using Simulated Dialogue to Build Confidence Through Role Play

One key design decision was to use dialogue between two avatars to represent the interaction between roles.

As discussed in the showcase, 

“that dialogue between the two people makes it feel like a real conversation, not just someone talking at you.” 

Learners are not positioned as passive viewers. Instead, they are invited to observe how each role approaches the interaction.

This approach supports empathy and perspective-taking. As noted in the session, 

“you’re building that empathy… this is a person in my role saying why they’re doing it.” 

Learners can see how intent, reasoning, and professionalism are expressed on both sides of the conversation.

Visually, the design reinforces the role-play dynamic. Rather than both avatars speaking directly to the learner, the layout creates the sense of an exchange. 

“It kind of zooms in on one, then the other, and then they come back together,” 

mirroring the rhythm of a real interaction.

Design Strategy 2: Designing for Iteration and Precision

Another critical design consideration was the ability to refine language.

In high-stakes interactions, wording matters. Small changes in phrasing can affect clarity, confidence, and tone. With AI avatars, these refinements are easy to make.

As noted during the showcase, 

“if you get those little wording tweaks, you can just edit it in here and then regenerate it and they just say it different.” 

This allows the learning team to refine scenarios over time, based on feedback or changing expectations.

The efficiency gain is significant. 

“You don’t have to go and re-record and then re-edit a film, which if you’ve taken it over an actual person, is a huge amount of effort.” 

This makes it practical to keep learning materials current and precise, rather than settling for content that is close enough.

Design Strategy 3: Choosing AI Avatars for the Right Kind of Learning

An important theme in the showcase was discernment. AI avatars are not a replacement for real people in all contexts.

As discussed, 

“I’ve never fully understood the value of a completely made-up person unless it’s a scenario.” 

In this case, the scenario is the learning. The avatars are not pretending to be real individuals. They are clearly modelling roles within a structured interaction.

This clarity matters. 

“It makes sense when it’s made-up scenarios rather than replacing real people.” 

Learners understand that what they are seeing is a simulation designed to support preparation and confidence, not a substitute for authentic organisational voices.

The guiding principle is intentional use. 

“It’s about working out the right user cases,” 

and role-based preparation is one where AI avatars add real value.

The Team Reflection: Changing Expectations of What Role Play Can Look Like

Team reflections highlighted how expectations around video and role play are shifting.

There was recognition of how far the technology has come. 

“They’re a lot better than they used to be. They were quite robotic at first,” 

followed by, 

“I am quite impressed with how far they’ve come in the last couple of years.”

There was also a sense of curiosity and possibility. 

“I still think of videos as filming real people, so it really fascinates me what we can do now.” 

Alongside this was immediate, instinctive feedback: “it looks really cool.”

These reflections point to a broader shift. Role play no longer needs to be live, awkward, or logistically complex to be effective.

Closing Insight: Using AI to Support Readiness and Confidence

This showcase reinforced that the real value of AI avatars lies in how they support preparation.

By modelling structured interactions between roles, the videos help learners understand what to expect, how conversations flow, and how to show up with confidence. They provide a safe space to observe, reflect, and rehearse before the real interaction takes place.

When paired with thoughtful use of role play and the ability to refine language easily, AI avatars become a practical and powerful way to support learning in high-stakes contexts.

Partner With Us on Scenario-Based Learning Design

Our team designs scenario-based learning that helps people feel prepared for important interactions. We use technology thoughtfully to model real situations, support confidence, and make learning practical without being performative. 

If you are exploring role-based learning and want to do it with care and credibility, we would love to talk.