How AI-Powered Role-Play is Transforming Soft Skills Training

---

Organizations often turn to soft skills training to gain a competitive advantage in the workforce. Andrea Laus, CEO of SkillGym, explains how AI-driven role-playing simulations are transforming leadership development, sales training, and customer engagement —turning practice into performance.

 

Show Notes:

Andrea Laus explains how AI-driven role-playing simulations provides real-time feedback, emotional insights, and data-driven coaching. His key takeaways include the following:

• Soft Skills Training Needs to Be a Process, Not an Event: Traditional one-time training isn't enough—improving soft skills requires ongoing practice and reinforcement, just like going to the gym. Consistency is key.
• Emotional Feedback Enhances Self-Awareness: SkillGym captures the emotional response of conversation partners, helping learners understand how their words and behaviors impact others—bridging the gap between intention and perception.
• AI-Powered Role-Playing Delivers Measurable Results: SkillGym connects training outcomes with quantifiable metrics, ensuring learners see clear progress. Studies show that employees who train with SkillGym outperform their peers in sales, leadership effectiveness, and customer engagement.
• Customizable Scenarios Improve Organizational Alignment: Organizations can create tailored role-play experiences that reflect their unique values, challenges, and communication strategies, ensuring training is relevant and scalable.
• Continuous Learning is the Future: Instead of training before work begins, Andrea advocates for learning on the go, integrating AI-driven role-playing into daily workflows for real-time skill application and reflection.

---

Transcript:

Susan Cort: [00:00:00] Conversations drive business, but how can training ensure that conversations lead to better leadership, sales, and customer service? 

Andrea Laus: The methodology is based on one single idea, that improving our soft skill is not an event, it's a process. So, we love to think about soft skill as hidden muscles that you have and basically you should treat them just like you do with physical muscles, which is going to a gym and sticking to a schedule.

That's Andrea Laus, 

Susan Cort: CEO of Skill Gym, a leader in AI driven role playing simulations for soft skills training. Find out how this type of training can transform workplace learning by turning practice into performance. Andrea talks with d’Vinci Director of Learning Technology Eric McDonald and me next on Powered by Learning.[00:01:00] 

Announcer: Powered by Learning is brought to you by d’Vinci Interactive. d’Vinci's approach to learning is grounded in 30 years of innovation and expertise. We use proven strategies and leading technology to develop solutions that power learners to improve quality and boost performance. Learn more at dvinci.com.

Susan Cort: Welcome, Andrea. We're so glad to have you with us. 

Andrea Laus: Thank you. Thank you, Susan. Very happy to participate and thank you for inviting me. 

Eric McDonald: Thank you as well, Andrea. Thanks for being here. 

Susan Cort: Andrea, let's start off by giving us a little bit about your background and then explain what SkillGym is.

Andrea Laus: Yeah, of course.

Well, I'm an entrepreneur and, uh, I actually live in Switzerland, but my company is, uh, international because we have, uh, headquarters here in Switzerland. We have an office in New York. We have an office in Italy. So we're getting somehow global. And, uh, I have a very, very good passion in learning methodology and neuroscience.

And this is probably [00:02:00] what brought me here. I've been studying these subjects since 20 years now, and our first attempt of providing a skill gym solution dates back to the early 10s, so we can actually say we are uh, Pioneers in this, in this area, much, much before the generative AI came on the screen. And, uh, I've been hiring a number of neuroscientists in my team, and we work with universities specialized in neuroscience because we actually have been developing on studies in neuroscience since the beginning.

And this is what our product is based on. 

Eric McDonald: I got to see SkillGym in action at DevLearn and see its impact it can have on learning. Can you explain to our listeners how it all works? 

Andrea Laus: Yeah, absolutely. Well, first of all, SkillGym is a digital role play platform. So the idea is that you log into SkillGym and you play a role play where you basically take the role of a [00:03:00] One person, which is typically a boss or a salesman or whatever else, and you deal with a counterpart, which is at any time.

It can be either your client or your report or whatever you have a challenge to overcome, and that goes through a conversation. That's what the role plays about, and there is no change with respect to the traditional roleplay. SkillGym is a solution that does a couple of things. First of all, we frame the way of thinking of best performers so the idea is that you practice on conversation being helped by the methodology of the best performers that normally address that conversation the best way possible.

And in this way the idea of SkillGym is let's take the abilities of best performers and make them available to the rest of us. To make sure that they practice in a smart way, which we call discovery and sensemaking or accelerated expertise [00:04:00] training. Basically, the idea of SkillGym is that of practicing in a smart way to save time and make it more efficient.

And at the same time, turn practice into a routine with a personalized schedule for users and also leveraging peer support because you practice alone, but you do that in a boot camp of multiple people, each one playing alone, but still connected to each other to exchange experiences and discussion. 

Eric McDonald: So one of the things that stood out was the feedback sessions that happen after you have a conversation.

Can you describe how the software is able to evaluate success and failure in something as nuanced as a conversation and the importance of emotional feedback afterwards. 

Andrea Laus: Yeah, yeah, great question. In fact, measuring something so subjective as human behaviors is not that easy, isn't it? What we do is we start from the very beginning, which is the [00:05:00] scope of that conversation.

So what we do is we frame the strategy behind the way a conversation should be addressed. And this is not done by chance, but it's done by framing the strategy in the mind of the best performers. So we identify a best performer, we frame their strategy, and we turn that strategy into very detailed behaviors of how that strategy should unfold along the conversation in terms of what the best performance would actually say and how they address the specific subject along the conversation to follow on with the right strategy.

Then we connect those behaviors to our competency framework. Most often is the competency framework or competency model or capability model of our clients. And this comes in a very structured form with clear elements of quantitative evaluation. So the competency model comes with how and why we can say that you're actually following a competence in the best way [00:06:00] possible, given the behaviors that you are applying.

So what we do is we collect the utterance of the user during the conversation. We connect them with the behaviors. We turn that into a measurable asset against the competency model. And at that point, we are capable of returning both a qualitative and a quantitative feedback, which is amazing because we can still provide a coaching style approach feedback like you did this and this is what happened.

And this is very much qualitative, but we also can tell you look in a scale of one to ten, you position that utterance at, for example, seven. Because of these elements are given your competency model, which is very important because in that way you reflect your context in your let me say frame of priorities and values when it comes to providing the feedback we actually provide three different layers of feedback the first one as you mentioned is the emotional one.

This is the way to [00:07:00] unlock in the user the need for getting a feedback when we end a conversation with typically believe that we did good whatever we do I mean unless we did a tragedy but normally we think we did good. And it's amazing to see how people don't recognize that what really matters is the opinion of the counterpart.

And so what we do is, we actually heaves drop the counterpart, so the character talking to a friend after the conversation is done and telling them about their feeling about the conversation they had with you as a user. This is amazing because you can't imagine how many people feel stuck. But the idea that they felt something and they did discover that the character thought completely the opposite.

Eric McDonald: That was super strong, that part. That, that was super impressive. 

Andrea Laus: Absolutely, especially because it's very precise. So it's not about, uh, you know, I talked to Susan or I talked to Eric and [00:08:00] I didn't like it. No, he says something like, at a certain point, Eric said that. And these made me feel like, or at a certain point, when I talked to Susan, she was objecting.

And so I felt that. So it's very specific, but very emotional, very subjective. So very hard to, to discuss because that's the opinion of your client, which is by the way, anyone talking with you. In a conversation which is negotiation conversation so it could be your report it could be your real client but anyway is someone you need to get a win-win situation and so it's a kind of a client in front of you and when they say this this kind of things i mean you can rebate I mean it's something they got it as a feeling and you need to you know you need to take that into account.

Eric McDonald: Their perception is their reality. 

Andrea Laus: That's it. And by the way, this is very interesting. I was just discussing with, uh, with our client, we have about this. One of the problems they had was, uh, you know what the real [00:09:00] problem when we run traditional role play. Is not engaging people is about convincing them that the best practice we're suggesting is the right one because they have their own opinion and they always believe that the way I do is okay.

You can't imagine how many people enter a training without real self-awareness about the reason why they are there and so it's always like this training is not enough for me or is too easy because I'm already 20 years old experience something like this. So the problem with the trainers normally is, uh, how can I get those people on board in terms of the validity of the best practice?

Well, the answer is let the counterpart talk. Let's have this emotional feedback. I mean, you can't object. In front of a client telling you, you know what I felt disengaged or I didn't feel listened to or whatever. So that's the key point. Then the second step of the feedback is very analytical. [00:10:00] What we do is we review the full conversation and you as a user now take a third-party standpoint.

So it means that you can sit back, relax, eat your popcorn and look at your conversation like if it was somebody else's. And with the help of augmented reality, we tell you, look, when you said this, the client understood that. And then he reacted like this, and then you did that. And then this is what happened.

And this is why that behavior didn't connect to the competency framework. You should, you should work for it. And so this is very interesting because you move from a totally emotional feedback to a totally analytical one. And so you can get in. And once you actually understand with the emotional feedback that it wasn't that good as you thought, then you have the possibility to go inside and investigate why that happened.

And of course, those comments are very much related to the competency model and the negotiation model that you were assumed to adapt to. So it's very, very precise. Last but [00:11:00] not least, metrics because you can't improve something you don't measure and so what we do is return everything we mentioned so far into numbers and so we tell you about your self-awareness in score about your confidence with the strategy in score - impact on the client in score, objective reached or not, yes or no, and things like that. These three elements make up the feedback session, which is the most important part of any role play. 

Susan Cort: This is just fascinating. I mean, when Eric came back from DevLearn and shared this opportunity to talk with you, I thought this is perfect because we really like that intersection of technology and education and also the methodology that you offer on top of things.

And of course simulated role play has been around for a long time and even simulated role play with AI is becoming more common these days. But yours strikes me as different because one of the focus on ongoing learning that gym concept versus a one-time experience and also about your methodology, which seems like it's more the focus than the [00:12:00] technology.

Even though you talked a little bit about that, but dive into the methodology a little bit more, Andrea 

Andrea Laus: Yeah, of course. What I presented so far is how one single session happens and that's based on two key pillars. The first one is that. Don't practice randomly so don't go inside the role play trying to make the best of what you already know and where your shoes and where the shoes of a professional in that conversation for the strategy and go through a discovery and sense making approach this makes the first and each session very efficient but that's just the beginning the methodology is based on one single idea.

That improving our soft skill is not an event, it's a process. So, we love to think about soft skill as hidden muscles that you have, and basically, you should treat them just like you do with physical muscles, which is going to a gym and sticking to a schedule. So, the idea, it doesn't make sense [00:13:00] to apply the classic one-shot learning approach to soft skills, because it's not about knowledge transfer.

Learn English for free www. English. com I'm sure you, you share with me the idea that anyone would have known those principle over and over again. I mean, they've been telling them over and over again, but it's not just by repeating those principle to people that they turned that into habits. To do that, you need to go to a gym and start practicing.

And SkillGym is just like that, it's a gym. You need a schedule, you need to sweat, you need to be consistent, which is basically the most important factor. We normally say, it's not about how many sessions you do, it's about how consistently you play. Just like in a gym, I mean, if you spend eight hours in the gym the last day of your subscription because you feel guilt, I mean, you end up to the hospital, 

Susan Cort: You know, Eric and me, 

Andrea Laus: you know, that's everyone.

I mean, this example sticks with [00:14:00] everyone because it's quite normal. People subscribe to a gym and they don't join and then they try to do everything the last day. Consistency in diet, gym and soft skills training is what makes the difference. And then of course we do this by supporting with a number of elements like gamification calendar driven appointments and support from peers on the same bootcamp groups again it just like once again going to a gym you go to a gym in courses where you help each other and stay motivated sharing energy.

And also sharing comments and things like that. This is what happens in SkillGym. You play alone, but you share that with people. And while you play, you get an engaged by gamification, you get stick to the agenda by calendar. Triggers that help you stick to that. You have things like benchmarking scores.

That's kind of full of feeling like, you know, somebody else. doing better than me or is doing more than me. And so I need to keep up. [00:15:00] That's kind of concept. 

Susan Cort: FOMO and training. I've never heard of that before. 

Eric McDonald: That's cool. It makes sense. Practice makes perfect. 

Andrea Laus: Yeah, absolutely. 

Eric McDonald: One of the things I saw at your demo at DevLearn was the ability for organizations to set up custom conversations.

Talk a little bit about how organizations are taking advantage of that feature. 

Andrea Laus: Well, yes. Of course, you can practice in SkillGym In two ways the first is to go through a ready available catalog of conversations because let's be clear there are some subject where you can't reinvent the wheel so I mean when we go on leadership most of the leadership conversation has been already addressed the right way so what we do is tell the client look.

This is a ready-made catalog. We can fine tune it on your competency model and you're ready to go and they're very happy with it. But when it comes to conversation like, for example, sales after sales or those conversation that delve deep in [00:16:00] their models, clients want to customize the scenarios and this is what you can do in SkillGym.

SkillGym is an authoring facility allowing you to turn your way of thinking into a scenario that you can scale. Two people for practice now what's the benefits here the first one is our clients can turn intangible into asset you can't imagine how many times we arrived to a client and we have them making order into.

I'm tangible best practices around the company so we had them collect those best practice from best performance in certain conversation and add the structure into a process, which is not always the case because when you go to a best performance and you ask them how do you solve that conversation what did what do they say they say well i just did it.

Because they are used to, I mean, they, they are not used to turn that ability into a process. [00:17:00] They just sold the conversation. So when we come into the picture, we basically bring that knowledge into a replicable process inside skill gym that turns finally into a role play. So the first thing is that they turn an intangible into an asset.

This is very important because for example, some best performer can resign, leave the company. And their knowledge go with them, but if you can turn that into an asset, not only you can basically help people practice at the same time scaling training, but you can also retain that that ability and make the best of it over time, even when the right people that somehow perfected them are gone.

Then you can standardize the, the way they spread these expertise. One of the problem is not just about collecting that expertise and turning that into an asset, but making sure that everybody else can use that asset the same way as anyone else. So [00:18:00] turn that into a standardized role play, help people and help the organization.

Make sure that people train the same way, with the same level of quality and the same level of expected outcome. And by the way, by scaling that you can help lower performance move towards the expected standard faster. 

Eric McDonald: And does the SkillGym have staff that assists with that sort of tapping the best people in the organization and turning that into training? Is that part of what you do offer as a service? Or is that on the organization to hire the right people to extract that training from the expert? 

Andrea Laus: That's a great question so we have developed the authoring facility to make sure that it's straightforward to do that job alone because we have very very simple questions that drive you through the process and the usage of AI in between helps with a lot of well suited.

Examples that we [00:19:00] let you know what type of information we are looking for at which level of depth and granularity. But at the same time, of course, we offer a premium service with our experts to support clients, especially in the onboarding phase, but quite often also ongoing by supporting them in doing part of that very delicate.

Activity of bridging the AI, as we call it, but in principle, they can do that their way and it's pretty straightforward. 

Eric McDonald: Excellent. So you've talked about it a little bit, but where do you see, uh, role playing having the greatest impact on organizations and individual learners? 

Andrea Laus: I would say that, first of all, soft skills is probably the ultimate competitive advantage the organization can claim over the next years. Whatever else is, is definitely destined to become a commodity. And so soft skills is what really will make the difference. Of course you can have a [00:20:00] spike of, you know, innovation in technology, whatever, but in the end how their organization will keep up and they will match your, your abilities.

And the only difference is really about soft skills in the end. So I would say, generally speaking, that role playing and practicing on soft skills is definitely the best way to maintain and develop your competitive advantage, which is always made by people given that I believe that certain roles can benefit more than others, especially those rules that either they are in the front in the front end.

So I'm thinking about sales or customer service. So those people whose actions and words can impact. The quality of the relationship with the clients and ultimately they can impact their loyalty to the brand. And second i would say those people that handle other people that manage [00:21:00] other people and are responsible for the quality of life inside the organization of other people and I’m thinking about managers.

But even between peers, you can have a great impact. I think about certain role plays we developed around the concept of diversity and inclusion, which is a very, very hot topic and where people have little self-awareness about the impact of their unconscious bias and behaviors. And so that's an area, again, where, you know, helping people to deal in a better way with other people, including their peers, can definitely improve the quality of life inside the organization.

And in the end, the quality of the performance of people, which is once again, the number one competitive advantage that organization have. 

Susan Cort: Andrea, can you share an example of how an organization has used SkillGym to improve learning? 

Andrea Laus: Yeah, let me provide you with a couple of examples. Of course, I can't make any names, but I'm sure this will be, [00:22:00] this will be very, very explanatory.

There was a insurance company, for example, that applied SkillGym to a specific unit of sales of some of their products. What they discovered, and this is a very interesting point, is that people that trained on SkillGym, as compared to people that didn't train on SkillGym, at the end of the training program, they basically We're selling almost double with respect to the people who didn't train with skills, which may look something, uh, unbelievable, but if you think about that and you, if you think about the, I mean, the comparison with sport, it becomes very, very, I mean, trivial because.

You can imagine you run a marathon and you compare with someone that trained for that marathon for months and you didn't and then guess what's gonna happen? I mean, you probably won't even complete that marathon. We are not that [00:23:00] used to think in terms of sport and muscles when we talk about soft skills.

And so sometimes you think it's about luck. It's about talent. It's about circumstances. It's about the way you or the product you're selling or the, you know, the price or whatever. But I would say that 50 percent of the recipe is about the soft skill you apply and you can train and you can practice and you can improve those soft skills.

So to me, there was no, actually, there was no. I was not impressed, which means that if you practice, you improve. One thing that is very interesting is that we noticed a very, very clear correlation between the improvement in the role play and the improvement in real life results. This is amazing. We have run over half a million sessions where one session is one person completing one role play over the last 10 years over half a million session is a lot and we have analyzed those data in a very precise way to basically [00:24:00] highlight a certain number of patterns and what we see is that when we have the possibility to share it, With the clients, what actually happens in real life after you practice, we notice a great correlation.

And if there is one point I would like to highlight as well, which is that of training on the go. This is another myth that we actually dismounted about traditional training. I'm sure you're aware that most of trainers out there think in terms of, first you train and then you go on the job. Like, let's take a training and then you're ready to go.You're good to go. 

Susan Cort: But you need to get the workers out there sooner. 

Andrea Laus: Yeah. But generally speaking, we believe this is stupid because it's much better saying you practice on the way. So just like you have a sales call today, then tomorrow you can schedule a role play sales call. Why not? You don't need to practice before you go.

This is not the Olympics where you practice for four years and then you have 10 seconds to win the medal. This is [00:25:00] more about, you know, you keep on, you keep on practicing while you go. And this is where you actually make the best of SkillGym because you can inject into real life, what you learn in SkillGym, and you can turn what you learn in real life into reflection for SkillGym. So this is a balance that works very well. Dismounting the myth of which is basically taken from the go to school first and then you get a job. Yes, that's okay. But this is just the beginning. Then the point is a job is a sport and you need to practice on the go. So this is one.

Another example is about, uh, is about leadership. We worked with a very large enterprise in services, and they've been practicing their leaders on skill gym. Part of the leaders, I would say 5, 000 out of 9, 000, and that was just because of cohort of learning. So there were people that were practicing before.

And then they happen to run a 360, uh, survey among [00:26:00] their employees to assess the quality of those leaders in terms of a certain number of KPIs, including the, their ability to set the right climate or the number of resign. And so the turnover. And what happened is that the people that trained on SkillGym scored much better than those that didn't do the training.

And once again, that's about practicing. I mean, there is nothing to be impressed about other than trying to understand why they don't apply this consistently to all the population. This is a very important point. But again. We are pioneering this approach, not just the approach of role playing, but the approach of thinking of role playing like a kind of, you know, uh, recurrent scheduled practicing approach.

Eric McDonald: Right. Continuous improvement along the way. 

Andrea Laus: Absolutely. 

Eric McDonald: That's great. So what's next for AI role play? Where do you see it evolving in the future? 

Andrea Laus: Well, this is a question I ask myself every morning when I wake [00:27:00] up.

Eric McDonald: The million dollar question. 

Andrea Laus: Yeah, I guess I'm part of the problem because it probably depends on me also. So, well, I would say two or three things. First of all, more and more personalized approach on the needs of the single individual. Well, Of course, we're working on that, but there is a lot to do. And, uh, the more you do on that, the better it is.

And, uh, the second point would be not just sharing the expertise of the best, which is one of the key pillars of role playing in a smart way, but also more and more aware of where the single user starts from and how you get to that best practice. As fast as possible and as easy as possible. So again, adapting the role play to the individual in terms of their ability to digest the feedback and to [00:28:00] sustain the rhythm and the speed of the schedule.

So we are working on that. So more individualized. Today, I would say that we have reached a very, very good point in terms of results, even if it's not that personalized to the single individual. And so we work more on cohorts rather than, you know, being, it's just like when you go to a gym, a physical one, I mean, and it's the difference between hiring a personal trainer or Getting up personal training for a small group.

It's just like that. They both work well. Of course, the first one is probably more. More careful about your specific needs. You will improve anyway, but you will feel more, you know, a bit more cared about. So we're working on that. And I believe this is one of the biggest challenges, not because of being difficult, but because of being something that the end user will appreciate.

Eric McDonald: Is there, is there any concern or [00:29:00] roadblocks or challenges with the personalization in terms of individuals sort of providing the software, the information it needs to personalize it? So in other words, my, you know, my digital avatar or my. Putting my personal information into the system so that it can know me it does that play a part in in the individualization?

Andrea Laus: Well that's certainly a challenge because once again when we talk about a there is always a balance between what you disclose and what level of personalization you want so of course you can have both.

You can be anonymous and get the best of personalization, but this is as real as in real life. I mean, if I don't tell you about my preferences, you will never know about me. The good thing is that we can segregate your information. And by the way, they won't stay there forever. So we can actually say that as long as that's useful to you, we can keep that information.

And when it's not that useful to you anymore, we will simply delete. [00:30:00] That information. So this is this is an important point. I would say that anyway, the system doesn't really need to connect who you are to what your preferences are. I mean, once again, sometimes we forget that there is just one single very small hook that connects the information to the personal information, which is basically those data that recognize you like you know your name.

Your email which is by the way the only personal things we collect and so all the record information is separated is connected via alphanumeric. Triggers and so I may know a lot about you but I don't know it's about you so you don't have to connect it. 

Susan Cort: Thank you very much, Andrea. This is all fascinating, and you are passionate and articulate, and I would say you've probably gone through SkillGym yourself to work on your soft skills.

We can't thank you enough for joining us and letting our listeners think about how [00:31:00] simulations can be used in an even more meaningful way. 

Andrea Laus: Thank you so much, Eric and Susan. It was a pleasure for me to be here, and I hope that this conversation will help somebody. Listening to have a better understanding of the potential of role play in general and SkillGym, of course, as well.

Eric McDonald: Yeah, thank you so much. 

Andrea Laus: Absolutely. 

Susan Cort: Thank you. 

Andrea Laus: Thank you. 

Susan Cort: My thanks to Eric McDonald, d’Vinci's Director, Learning Technology, and our guest, Andrea Laus, for joining me today. If you have an idea for a topic or guest, Please reach out to us at PoweredbyLearning@dvinci.com. And don't forget to subscribe to Powered by Learning wherever you listen to your podcasts.