AI Inside Immersive Training

Gene Feldman: hello, welcome everybody to our next episode of Neon Protocol. I'm Gene, and we also have part of our Neon Protocol team members. We have Damon, Dan, Mike, and Ron. And with us today, we have our special guest from Virtual Speech, Sophie Thompson, who has actually an absolutely fantastic story of how she started and what got her here. And we're we've been really excited to talk to her. Your Sophie, your name's come up a couple of times in the past couple podcasts. So yeah, we're really excited to have this conversation. So if you want to introduce yourself to the audience, please.
Sophie: Thank you. So yeah, I mean my career has probably taken a different path to most of your guests. So my background isn't in tech or engineering or anything technical. I I come from a very much like a humanities background. So actually at university I studied theology, so very different to to what I'm doing now. And yeah, working in tech was never the plan for me and yet. my whole career has been in immersive learning. so how I got started on this path was when I was at university and I was studying for my master's in international business at that point. I had one of those moments in life where you're at a crossroads and it turns out the direction you take then determines your whole future. So it was 2015. I how old do I have been? 22. and I had severe social anxiety at the time. So I wouldn't even order my own food in a restaurant. for one of my like oral exams, I had a panic attack before because I just didn't want to speak. and I was very studious. Like I'd always loved to learn, I've always like put academic achievement quite closely tied with my identity. and I had an assessed presentation coming up, so you can imagine.
Damon: you
Sophie: the state of anxiety I was feeling. So I was waking up three months before this presentation, already nervous. And I was talking to my friend Dom about it. And Dom was working at Jaguar Land Rover, the car company, with virtual reality there. And I hadn't really heard of virtual reality back then. And Dom had the brilliant idea that VR could provide a psychologically safe place to practice a skill like public speaking in a safe, controlled environment. And so Dom, being the genius that he is, built the first version of what became known as virtual speech. And it was a way to help me prepare for my presentation. And so we ended up being the first virtual reality app to overcome the fear of public speaking. And it was really just set up as some nerdy fun to to help me. and within well, within about s six months we had a hundred thousand downloads. Within eighteen months, I went from not being able to order my own food in a restaurant to being interviewed live on BBC World News. So I'm a walking talking case study in that respect. so obviously a lot of time has has passed since then. but we have continued to be the first in in many respects. So in 2018, we built our own learning management system so that our clients could see data from the VR app because no learning management system could do it, so we built it. in 2022, we were the first XR training provider to integrate with GemTive AI because there was a real need for it, which we'll come on to. and yeah, we've worked, we've helped nearly a million learners worldwide. across organisations like PwC, Bayer, T Mobile to not just improve their skills, but their confidence to use those skills.
Gene Feldman: Well, I mean, j honestly, just hearing you give your spiel. I mean, Sophie, I've known you for years and the last thing I was said you had trouble public speaking. So I mean it Yeah, I mean it's a huge testament to what the software does. But now so out of curiosity, so w I mean it's a great origin story. now my one of the questions I have, how when you first released it, how many scenarios did you have and what was the time? Cause I cause I know where you're going with the genitive AI.
Sophie: I know you can't shut me up, Jean.
Ron: haha
Gene Feldman: on why it was so valuable. So I'm kind of keying it up to you know, before you had your LMS, did was it like one just practice scenario? Did you have different environments? And how long did it take to to build it? Did you have to branch it? So I'm kind of going in a couple of different directions.
Sophie: Yeah, that's a good question. So we I think we had the conversation probably December twenty fifteen and the first version was released like early February twenty sixteen. so it all happened quite quickly. I remember and for years we had environments in there where we had like green screen recorded our friends. so for years like the audience people were practicing to were our friends at the time and we just
Damon: you
Sophie: brought them all to like Dom's house, just put a green screen up and we recorded them. So the first version was like green screen recorded people on like 3D environments. And actually one of those 3D environments is still in there today and it's the most used environment, which is interesting. No, although I do know you're a fan of that one. It's the it's the big purple, it's a purple
Gene Feldman: okay. Is it the elevator pitch? I am a huge fan of the elevator pitch.
Sophie: Conference room. I say purple because there's like purple lights in there. but yeah, the first version had that room, a meeting room, I think like job interview practice. We also at the same time, or like in parallel, actually also built a language learning app, but we had to like pick a path, so we fogged that one off. but yeah, and we did very early have AI in there, not generative AI. But we had IBM Watson plugged in so you could still get feedback on your delivery. So pace, volume, how many filler words you said, how many ums and ahs. So that's always been in there. even like in the first version in twenty sixteen.
Gene Feldman: Nice. So now how is how is the genitive AI and all of that coming into play into these scenarios? Because I cause one of the questions we were kind of talking about before the podcast. Sorry, audience, you missed it. is we were kind of just generally talking about how it seems that AI is getting more of the tension, XR is kind of falling back, but I think your use case is a good argument for the blend of it, of the two. So I guess how does your AI fit into that now? Or the new versions of AI?
Sophie: Yeah, so I mean it's it's very much integrated into what we do now. so if I take it back to like the beginning, so I think it was November twenty twenty-two that Chat GPT like exploded in all of our faces and just erupted into the world. and I remember like everybody seeing like articles about it and thinking, wow, this is really cool. and again it was Dom who Messaged me and he was like, This could be brilliant for us. And I was like, I believe you, but what do you mean? And he said, This is the answer to our most requested feature from our clients, which was for the avatars to understand what they were saying, understand the context of what they were saying. Because we had environments, people to practice in to evoke that emotional response. We had feedback on the delivery, so how they've said what they've said, but we didn't have the ability to do feedback on what they'd actually said. so within six weeks, we had that live in the app. And that's that was like the first iteration of it. So that has really leveled up what we can do. It's taken it from practice and I I guess slightly more like static training to very dynamic. personalized. we've got over 50 role play scenarios now. So you can practice 50 conversations off the shelf or you can author your own now as well. And obviously the avatar understands what you're saying from a learning point of view. It is personalized to each individual learner rather than being kind of stuck in time in with what we have built at that moment in time. It adapts. So it's it's applicable to people at every stage of their career, whether you're Like a new manager or you're an experienced manager, the software a is able to adapt to them and give them valuable feedback based off proven theories and rubrics that we've put in in the background, but they get a far superior level of feedback now.
Gene Feldman: Very nice.
Dan Meeks: you mentioned this authoring. What what does that look like?
Sophie: So we have an authoring tool called Role Play Studio. And it's an online tool, but then you can create your own role plays that then are pushed into the BR app, but also online as well. So we have a lot of learners who access us via online via WebGL. And it's it's effectively a form that people just put the learner's role, the avatar's role, the context for the learner, the context for the avatar. So effectively like the avatars. brain, but it's only needs to be like a couple of lines, doesn't have to be loads. And then if there's a specific framework you want the feedback to work with. So if say you want it's an onboarding and you want people to learn about company values, you would put your company values in that feedback section. And then you just click a button and it generates the role play. It generates five key skills for the feedback criteria. So in the example of company values it would take five of those. and it's all outlined like what that feedback would be based on, and you can edit it, but you can have the role play live in, I mean, literally two minutes, however quickly you can fill in those sections.
Gene Feldman: And so based on this, just for our listeners, so it it almost sounds like there's no branching. So there's no more A, B, C, D answers. It's you just get to talk naturally to it or type natur however the input is. But rather than a multiple choice assessment, which we always know puts people in a binary because people always say, Well, I would what I would do is not there. I I I
Sophie: Yeah.
Gene Feldman: I would do something d completely different. And so I think having that natural input, so is that kind of how it's set up?
Sophie: Yeah, that's exactly how. So there's absolutely no branch learning. There's no kind of way to like click back and then cheat the system because you know you're supposed to do XYZ. yeah, it's entirely responsive to the learner's unique input.
Mike Vann: Sophie, quick question. there there's so much p possibility here. So it's really great what you're doing. we work with enterprise, so we we know some of the issues around this, but wondering how when when you speak with companies and they want to do this type of thing, I know they get concerned about what the employees might say into the AI. You know, what what where that data goes, whether or not it's released into large language models. I mean, have you have you
Ron: you
Damon: you
Mike Vann: Come up with tactics around that.
Sophie: Yes, I mean, you're totally right. That comes up with every enterprise client. so we have a paid version of the LLM we use. and we have a contract that says that it's not being trained on their data. We also isolate the data across each company. So it's not like say if PWC are using it and Bayer are using it. It's not like they're then feeding our unique system, like it's it's isolated across the board there.
Mike Vann: Yeah. Yeah.
Sophie: And then we also have guardrails in place. So for example, if somebody set uses like racist, misogynistic, homophobic language, it will just kick them out of the system and it will come up with it will tell them the reason why, to also protect from people just not being very nice on that.
Gene Feldman: Very nice.
Ron: You have demo, salty of what you do.
Sophie: Do I have it I can send you a demo, yeah, absolutely.
Gene Feldman: No, I know. I
Lucky Gobindram: Sophie, is it also desktop at this point or is it only VR and MX reality?
Sophie: No, so it's on desktop two, it's been on desktop for probably three years now, maybe longer. But yeah, it's on desktop two to to address that scalability problem that XR has.
Lucky Gobindram: Do you see that as the gateway to opening up enterprise conversations? Do they usually start desktop before they actually go into the the spatial versions? And do they do you d are they like what percentage of of this is actually deployed to spatial versus it just being de a desktop application?
Damon: That's a better question. The other one was leading the witness.
Lucky Gobindram: Yeah, yeah, yeah, yeah. That's why I I thought I rephrased it quickly, didn't you?
Gene Feldman: Ha ha ha
Mike Vann: Yeah.
Sophie: so about 50% of our clients use XR, 50% are desktop only. Our biggest client is an enterprise client in the pharmaceutical space, and they use it with nearly 25,000 learners across the world. So obviously XR is pretty inaccessible from that point of view. So they use the the online version, and that's where I see the real scale with. what we do is having it available online. We originally developed the WebGL version for for exactly what you said to kind of as a gateway, like a stepping stone into XR. but we haven't actually found that it has acted like that. If people start with the online version, they tend to stick with the online version because the problem isn't that they're not convinced that the XR solution is works and is the better one in terms of that emotional response that you have to it. It's it's everything around it. It's the logistics of the hardware, it's the cost of the hardware, it's convincing other people that XR isn't just this gimmick or gaming and it's it's it's everything else around it that is that we find is the problem.
Lucky Gobindram: Sure. There's a there's a lot of noise around it that you have to like silence or or overcome to to actually get a pure spatial deployment that's focused on this. Can I ask you one other question? So when you're in there, right, and you're you're doing it, it's real time. So like there is there any like input that you have to trigger, like say I'm having a conversation, right? And a lot of the ones I've tried before, they they they make you have like hit hit a button till I complete it and then like it takes the input and it goes. Is this
Sophie: Mm.
Lucky Gobindram: purely free form, like real time.
Sophie: Yeah, so it's in real time. We actually had to slow down the response time. because people who were either using the software in a second language, so if they're from another country and they're using in English and needed that moment to think, the AI would interject. Or if you just needed a moment to think in general, it would assume you finished speaking and then it would respond. So we actually had to slow down the response time, but you still don't have to click. a button to to say you finished. and I would say the I mean I can have a send you a demo, that you can see that the response time is very realistic.
Dan Meeks: That's a true thing.
Lucky Gobindram: I I d I've tried your I've tried your tech a long time ago. So I ha I've i I've interacted with it, just not not recent.
Sophie: Yeah, I mean you might have interacted with it when when we were using what we now call manual mode, where you have to click to say you've finished. we have a couple of clients who do actually prefer manual mode, to give people that time to think, but most of the time it's much more realistic for it to be able to have a conversation like we're having now.
Lucky Gobindram: Okay.
Dan Meeks: Now can you take that dial and move it the other way? Because as an engineer I sometimes people will pitch me with an idea and I get really, really excited and I end up talking over them, which probably makes them nervous. But I'm just thinking a little bit faster than they are. so I I mean it probably doesn't happen very often, but could you take that dial and go the other direction where you're not waiting, you know, to to actually increase the stress during the training session?
Gene Feldman: Yeah. HE
Sophie: Yeah, I'm trying to think what we did with this because at one point we did have avatar interruptions. Then we removed them because people didn't like it. and then we had a discussion about adding them back in. I actually can't remember where we've landed on that. I think at the moment the avatar doesn't interrupt people because they would get frustrated and say the avatar didn't let me finish.
Dan Meeks: Could I do that inside the authoring though? Could I could I actually say, Hey, the the person I'm talking to, the person I'm giving the pitch to is known for being a fast talker and a quick thinker and you know Exactly. Yeah.
Damon: An interruptosaurus is what, yeah.
Gene Feldman: Yeah, al almost almost set up the personality of the avatar to mimic what you're going to experience in real life.
Damon: of the, yeah.
Sophie: Yeah.
Damon: Dynamic AI.
Lucky Gobindram: For sure, isn't
Gene Feldman: Yeah.
Lucky Gobindram: that isn't that interesting? Right. Like I I mean, I think it's more relevant honesty guys in when you're in spatial, right? And if you think about it 'cause like it's like really like bringing a believable human into the the into the environment, right? two D it limits it. It's like you're talking like we are right now in Brady Bunch boxes. So like it's a little bit more behav like you know, you expect something, but like you expect a human when you put somebody in front of you, right? And like it it becomes like you have to they have to have the interruptions, they have to have the pauses, they
Sophie: Yeah.
Lucky Gobindram: it's not believable, right? It kills the it kills the believability.
Gene Feldman: Yeah. I mean like that's us.
Sophie: So you can put that in in the Avatar persona. sorry.
Gene Feldman: nice. No, no, it you and you get straight to the question because I was actually just gonna become say like the reason why I love the elevator pitch one so much is because we stress the elevator pitch, you know, it's 30 seconds. But when you're in an elevator and you're seeing the floors click by, you know, that all of a sudden puts that timer right there in front of you. You know, and just the preparation that that can give and being able to set that avatar to where, you know, they are interrupting. They're looking at their watch or, you know, they're just I guess not engaged is I guess where I'm
Sophie: Yeah. And they do have reactions. So we have eighty different like reactions and animations in there. So that for example if you if you are rude to the avatar or make the avatar angry, they'll like cross their arms, they'll frown, they can cry, and if they get really angry, they'll like slam their fist on the table. And yeah, so they can do things like check the time or if you built it into role your role play studio
Gene Feldman: that's awesome. Now
Sophie: custom role play, when you put in the like avatar's brain, that's when you give it its persona. So you could say you interrupt after thirty seconds or something.
Gene Feldman: Nice. Now speaking with your avatars, so one question that I know always comes up in the XR world, you know, and now with the AI world is the uncanny valley, you know, that we want to go for this heightened reality, but we want to avoid cats. You know, it's you know, we we don't want it to go too far. So how I guess how have you I guess navigated that? But in the midst of you really came at this from outside the XR world. So how did you navigate having to I guess this is kind of a two part question, navigating, learning a lot of this and then really implementing to make sure you're not hitting these key areas, such as the Uncanny Valley? Did that question make sense or did I just I kinda add it multiple in there at the same time?
Sophie: Until It made sense, but I may forget the second one by the time I've answered the first one. So you may need to prompt me. in terms of like how we've developed it, it's very much being centered around what our customers want and our customers' feedback. So we have like a giant spreadsheet, and every time we get any customer feedback, we add it to the spreadsheet, we say which customer, and then we go through that, and that's what is decided how we how we build things.
Gene Feldman: Yeah.
Sophie: Our key customers and our most vocal customers have really shaped the product to be what it is. In terms of the Uncanny Valley and what the avatars look like, so you will see, or anybody listening who looks up virtual speech, will see that the avatars aren't photorealistic. they're not trying to be. They are emotional enough and realistic enough. We have clients who, especially in the last like year or so, Our competition with enterprise isn't other XR companies, it's AI companies. and or other AI learning platforms. So, for example, we work with a big insurance company and they did a pilot with us, and at the same time with another group, they did a pilot with using like synthasia avatars. Obviously, they are photorealistic avatars. But actually their feedback for that was very resistant. And their employees felt like the company was trying to trick them, was cheaping out by using AI and was trying to fool them into thinking that it was a real person and it wasn't. And there was this huge resistance because many of the people on the pilot felt like they were being tricked or trying to be tricked and didn't then trust. not just the learning, but the company as well. So it like damaged that trust, like feedback loop with the company. and that's why they went with us, because they didn't have that negative feedback around the avatars. Everybody entered knowing it was an avatar. It wasn't trying to be something that it's not.
Gene Feldman: Yes.
Lucky Gobindram: Super interesting.
Ron: kind of benefits do you highlight, to show your advantages?
Sophie: Our unique selling point really is the feedback. So we have an up an ad adaptation of the experiential learning cycle. So it's practice, feedback, reflect, apply. So you say have your role play conversation, that's your practice. You get that detailed feedback, and then you speak to a one-to-one like AI careers coach about your feedback. You get more insights into it, you can ask the coach questions. You can ask ask it for examples of where you what you could have said instead. and those are the I mean, to be honest, the most important things, but also the the reason that people tend to go with us in terms of results. So we track a lot of data points that are all available to to clients on their admin dashboards. And we have seen a Average skill improvement of 25% in two sessions and an average increase in confidence of 28%. and when you think like so, we also work with higher education, and if you think like if you're say like 20 years old and you use something that improves your skill by 25%, your confidence by 28% in public speaking, job interviews, whatever that might be, that then has a compounding Effect as well. And that is life changing. If you learn those skills at such a young age, that then sets you up.
Ron: So you have examples of these synergies really fast tracking a person's career or making much more money or being more strategic in their thinking or what?
Damon: you
Sophie: Yeah, so we sometimes have people who will contact us directly. and then obviously our clients will feed things back. my favorite stories that I always tell people actually on their onboarding on their first day with us are the the more conversational stories that people just emailed and told us. So we've have people, somebody who contacted us who said that they got a job at the UN because they used virtual speech to help them practice for their interview. we have stories of people say like raising money for their startup because they practiced their pitch. then we also have I would say more of like the feel good stories are my like favorite stories where there's that social impact. We had last year somebody contact us who is working with victims of trafficking and they are using virtual speech to help them prepare for interviews and explaining gaps in their CV. We have an organization in Sweden that works with public services in a a big region in Sweden and they're using it for crisis communications. They use Roleplay Studio and it started off as speaking to victims of natural disasters or terrorist attacks. And then this year it evolved to speaking to In hospitals to victims of domestic violence and mental health first aid, so suicide prevention effectively. And only I think it was about two months ago, I I spoke to that client and she said, like, Sophie, virtual speech is saving lives. And I was like, well, that's good to hear. And she was like, No, really, it is. And she was on her way to do some of this training when she had her headset in her like headsets in her bag. And she was waiting for her train and there was somebody at the train station who was was going to walk, well, walk in front of a train. And she said she used the training that she had done with virtual speech to speak to this person and like brought them to safety, basically. And I mean, I don't know if anybody in the room or or listening has ever done like psychological first aid, but In that situation, a lot of it is around negotiation. So for example, people might say, will you feed my dog for me? Or will you tell my parents I love them? And every part of you wants to say, Yes, of course I will. But that's not actually the the right thing to do to prevent the the situation from escalating. You you have to actually stay quite firm and you have to say, No, I won't. You will do that. so it's it's obviously a very
Gene Feldman: Mm-hmm.
Sophie: complicated and complex communication skill to have. And I think that really shows not just the impact, but the the breadth of capability and the breadth of the communication skills that there are. Like it's not just about public speaking or having a difficult conversation with somebody on your team. It it goes into every area of your life.
Ron: for us.
Mike Vann: So the so the the feel good stories are tremendous. Those are amazing and I'm sure they help when you're talking to people about what the platform is. so I'm gonna bring it all the way to the other side. So as time goes on here and as kind of this push a little bit of pushback on AI from people, maybe people that don't understand it, or maybe people that do understand it, are you starting to see, you know, some small percentage of people that come into this in a skeptical way and try to trip it up or tried or are concerned about whatever they're saying to it might be used against them and their company. I mean, I know some of that is happening and there's always a small percentage of learners or people that are using a tool that are that are gonna have some issue with it, do you hear from them?
Sophie: Yes. I mean people try to break it all the time, but usually it's in like a a f a fun way. But only three weeks ago now, we did get some feedback from there's a company using it across like different core groups and one of them is in a a distribution centre in a warehouse and we received some feedback from them.
Mike Vann: A fun way. Yeah. Yeah.
Sophie: And they basically said, have you had this before? Do you have any advice? And I opened the feedback. And with the time difference, I was lying in bed. It was like 10 30 at night. And I opened this feedback. And I was like, my God, like this is, this is a problem. And there was nothing in there about the platform. There was like obviously there was good things, but the you learn more from negative feedback than you do positive. So I like hone in on the negative.
Mike Vann: It's
Sophie: And the negative feedback was nothing to do with the platform, it was about AI and resistance to AI. And so yeah, to your point, I think as the general society sentiment has shifted from wow, this is so exciting, look at all these things we can do, to you know, massive companies doing layoffs and blaming AI for that, even if AI isn't the real reason, it's a really good scapegoat at the moment. And so this general sentiment around AI is shifting. People are more suspicious. obviously, with with that example, there's the population that that the training was done with are typically more susceptible to being or scared of being replaced by AI. but there was one person in there who it was quite clear that he thought he was being assessed by doing this training because he He got something he got something very wrong in the conversation. And then the AI obviously picked up on it because it's acting like a human. Yeah, and gave feedback. But this person was very upset with their feedback and was really trying to justify it and getting really angry with the AI because of it. And like underneath, I mean, ironically, the AI system did exactly what it was supposed to do in terms of pulling someone up on something they shouldn't have done. But underneath anger.
Mike Vann: Right. And gave feedback. Right.
Sophie: is usually fear. And I think it it came from fear. So our solution to that. Yeah. Yeah, exactly. And up until that point, we had always had people actually more responsive to AI feedback than human feedback because people didn't get offended by AI because it was seen as this more objective thing. It didn't have you know, it was less awkward to give
Lucky Gobindram: He took it personally. He or she took it very personally in that moment.
Sophie: constructive feedback, people didn't get offended because the system didn't have feelings, etc. But that was the first time that I thought, ooh, the sentiment is starting to shift here. And I mean the s the solution, I I ended up obviously spoke to them. I sent them like a four-page document about the importance of implementation. And I think that's going to become increasingly important with any AI systems that are used for learning is ensuring that they feel
Mike Vann: That's right.
Sophie: psychologically safe to then use it, that it's not being used to assess them. And we would never advocate for our platform to be used for assessment. I mean, in the EU, for example, that's literally not allowed. But also just from a I guess like a moral point of view, like I don't think that it should be. It's there as an aid. So we have like universities who will use it and people
Mike Vann: Safe. Yeah.
Sophie: say deliver their presentation in virtual speech, the admin can see the feedback scores, but they can also see the transcript, listen to the audio, and the scores aren't there to replace the the human input.
Gene Feldman: Now I want to kind of go on a little tangent here because you mentioned something as we're talking about AI and the pushback. I think one of the things that a lot of people kinda not want to say forget about, but it kind of gets caught in our blinders is governmental interaction with it. You know, and you said something that there's the the there are certain things There are certain things. Apparently English is not my first language, there are certain things that
Damon: you
Gene Feldman: you can't do over there. And like I know a couple of years back there was the EU AI Act. S and especially us and us colonists over here in the States, we often forget that there is an impact between what happens globally. So what are some of those limitations of what you can and can't do, I guess, versus the states to what you have to in interact with at home?
Sophie: That's a really good question. It's quite difficult to answer because I'm based in the UK. I mean, most of our clients are in America, but because I'm based in the UK, and unfortunately, we're not in the EU anymore, but that's another topic. But you know, we're very close proximity to the EU and we are influenced by EU laws. We by default build for those laws, like with GDPR. It's more stringent and I mean better for the end consumer than laws in
Gene Feldman: Okay. Ha ha ha.
Sophie: the US are. So actually the there tends to be a higher standard because of these EU protections. So by building for that, we then essentially cover ourselves for laws in the US. I know like California is more strict, for example, but it's not more strict than the EU. So that's that's good in that respect. But in terms of like the the EU AI framework, that's there's various things in that and it and it keeps being added to so say when you upload virtual speech, like you will see like there is something that says this is using AI, this you know, it can make mistakes, so you have to almost have those disclaimers. but there's various things in there about I mean, that one for example covers off like having to be transparent about using AI. when we're like talking to people, we we s you're not allowed to use it for assessment. So we will s if somebody comes to us and says, we want to use it to to assess this person and it'll be their promotion will be based on this score. Like morally that I don't feel comfortable with that. But also legally in the EU, none of our EU customers are allowed to do that. And they don't know that they're not allowed either. I once was at a an event last year and there were other XR providers there and somebody came over to me and they said, I've just been talking to X Company and we're going to use it for assessing students for XYZ and these our employees for A B C. And I was like, that sounds great, but you're actually not allowed to do that. and they didn't know and I assume the the company didn't know either. So I would say there's more processes to keep up with over here, but it's it in a way just means we're like a step ahead of of when the US gets to that point, generally speaking.
Gene Feldman: absolutely. And I just remember a couple of years back, right before Copilot, there were a couple of companies that were like creating sales tools that would track sentiment. You know, like as you're talking, it would use AI to track the facial expressions to identify whether people were engaged or this or that. And then all of a sudden the EU AI Act came out, which banned all of that. And there were companies that weren't paying attention that were still developing it. And it's realizing that. If it's happening over there, there there is a chance it will come here. And I think not acknowledging what's happening on the different continents can actually be to somebody's disadvantage. And that's why I was k really curious about that question. I mean, I don't have a follow up question, it's just a follow up statement.
Damon: Well, I have a follow-up question. Well, you know, the various guardrails and guidelines around AIXR, ARVR, all those sorts of things and assessments. I mean, it does beg the question, oh, well then what metrics can you keep? What metrics can you measure and what can you use them for? And specifically something you mentioned early on, Sophie, is that you were measuring
Gene Feldman: Please.
Dan Meeks: Right.
Damon: Confidence and I was really really curious How does one measure confidence is that like a subjective? I mean, what is it? What how do you define it? How are you measuring it? And then how do you stay abreast of the laws to make sure that you're not taking a metric and using it in some way that you shouldn't be That's a lot. Sorry
Sophie: Yeah. No, it's fine. I j I don't know if you saw me typing, I just wrote quick notes then. so in terms of the metrics, it's not that the metrics can't be used at all. It's that they can't be used to effectively like impact someone's career. So you can't use the metrics to say that Damon is more ready for this promotion than Dan. Like you can't use them in that way, but you can use them to
Gene Feldman: Ha ha ha.
Damon: Hahaha!
Sophie: To see like improvement, to see where they're at in that skill set, to see the benchmark across the cohort. And crucially the end user can use them themselves, but from like an ROI point of view, they can still use those metrics, apart from companies like Germany, where they're not allowed to see them for each individual user. So, yeah, in which case they're not allowed to have access to the admin dashboard. So, like there's a good example. Yeah, exactly.
Gene Feldman: Interesting.
Damon: Wow. Well, that would have prevent misuse, right?
Sophie: so but generally speaking, like all the metrics can still be used. They just can't be used to make a life changing decision, basically. and then in terms of the confidence, that is is based on not one of our metrics within the app that is based on like feedback surveys. so asking them before to to rate how confident they feel and then asking them afterwards to rate how confident they feel.
Damon: Okay, so that is sort of a subjective, yeah, gray scale, really a gray scale. Yeah, metrics always, that topic comes up over and over and over in XR in the hardware response and the software response. mean, everything, latency, emotions, all of it. It's our favorite topic.
Sophie: Mm, yeah, yeah, yeah. Yeah. Yeah. And I mean things like percent improvement. So I mean, we collect over, I mean, in total, there'll be over a hundred data points. that I mean, the admin can see them all if if they would want to spend their time going through that. But they also get some top level metrics, and one of them is percentage improvement, and that's like looking at like their first score compared to their their latest score.
Dan Meeks: Well, you you mentioned earlier that each client has their ability to pick their own AI model to to kind of stovepipe the data. But I can I could foresee some companies wanting to prevent any kind of indemnification by basically hitting the flush button every once in a while just so that data's not around so they don't get roped into a lawsuit that might have something other to do with the data. You know, just a way to kinda keep their hands clean on it. I assume. But is
Sophie: Yeah.
Dan Meeks: That's the case.
Sophie: Mm-hmm.
Damon: Is that a thing?
Sophie: What people will sometimes do if they are in, say, the healthcare space or banking is that they will request for some features to be turned off. So they won't have the feature of say seeing the learner's transcript will be turned off and the recording will be turned off. So learners can't you always have to opt in to record yourself or save and send the transcript, but Just to kind of cover their backs, they remove that option for the learner altogether. And sometimes they will also use anonymized accounts. So it can't be linked to an individual. But where that comes in, this is a GDPR thing for you, is that AI has opened up a bit more of a can of worms around voice data. So even if the account is anonymized. because the voice is being used and even when we're not saving it, because it's having that input, some people still ask for that to be turned off because it is still a protected data. Even though it's very unlikely to be able to identify one specific person in the world based off their voice. I mean, maybe in the future it won't be that difficult. So then it can be turned off.
Damon: It appears to be becoming less difficult. Maybe put it that way.
Sophie: Yeah. This probably this comment probably isn't gonna age well. If anybody listening to this in two years might be like, gosh, the naivety of twenty twenty six.
Gene Feldman: I was
Damon: Two years, maybe two months. I mean it may be like August and someone's going, this is out of date.
Sophie: Ha ha ha. Yeah.
Gene Feldman: No. Well, and I can see it almost the using it to determine is this an AI recording of a person's vo of the person's voice or is it the actual person they're talking to? You know, to avoid the deep fake.
Sophie: Mm-hmm. Yeah, yeah.
Damon: Yeah, yeah. The forensics of the data, yeah. Yeah.
Sophie: Yeah. I think that's why offline is actually going to become more popular. I think we're going to do a bit of a full circle there, that people become less trusting. I think people are already less trusting of digitally what they see on, say, social media. But even in terms of calls, I think people become less trusting of is that the real person they're speaking to? Like, am I being recorded? I've been on a conversation with one of our clients who's a a law firm, and I didn't have my camera on for some reason and like they asked me to turn the camera on to make sure that it was me and it wasn't an AI. so I I think there will be, yeah. Yeah. Yeah. So I think the the off offline will become more important, like in person meetings, in person events to build that, to have that trust, but also I think we will crave that human connection more and more as well.
Lucky Gobindram: Are you human? Please identify yourself.
Damon: There's a Battlestar Galactica cut.
Sophie: So ironically, we're using AI to help people build human connection.
Lucky Gobindram: Pretty cool, you know, how I can reverse it, reverse
Damon: Ironically.
Ron: How do do that?
Lucky Gobindram: it back.
Dan Meeks: yeah. I love flying a sim, but man, I really love flying. So Yeah. Yeah.
Damon: The real thing, yeah.
Sophie: Yeah.
Ron: So Sophie, any insight on how we're technology to build relationships and build trust?
Sophie: gosh, that's a really good question. I think a good example would be a university that we work with, an American university actually, and they're using it for for various different use cases, and one of them is around student housing and part of their use case was conflict resolution in student housing, but also working with students with like special educational needs and An example of like building human connection as a result of using virtual speech is that one of their students with special education needs wasn't in student housing. And then they built their confidence and their skills and communication skills enough with virtual speech that then the next year they did choose to go into student housing. and the our client reported that back as a direct result. that same client has results around. they use the job interview preparation. And I think it is something like 92% of students now that go to the career center ask to use virtual speech because it's helping they've heard how it's helped other people with their communication skills.
Mike Vann: Sophie, we've been talking a lot about AI wearables and the kind of next the next gen of hardware for XR and glasses specifically. So what are your thoughts about some of this and how glasses and audio come into play? I mean I I can imagine and I know that enterprises are already, you know, can getting eyewear i out into the workforce. is this the kind of thing that people can be walking into meetings and practicing before they they pitch?
Sophie: Yeah, I mean there's there's an obvious use case for virtual speech with the glasses in terms of effectively having a coach with you at all times and instead of having to upload your screen or put on your headset, you just tap and you can practice like directly in the flow of work, even in the conversation, after the conversation. so it's definitely something that's on our radar. My hesitation is I is the commercial viability of it. it's definitely something that we could create, but how responsive are people actually going to be to that? Because There's a lot of concern, say, from like a data privacy point of view or data and privacy point of view, and especially a privacy point of view. we actually had somebody on the team today who was talking about how they were viewing houses at the weekend, the moving house, and the estate agent had like meta-ray vans on, and they felt really uncomfortable with the fact they had them on. and I think We perhaps have all been in situations where we've noticed somebody's got them on and even though it doesn't say it's recording, there might be like a shift in you somehow of but are they? Or if they're not recording a meta recording, or it doesn't quite feel like you're the only person in the room suddenly. And I think
Dan Meeks: No, I'm just a jerk about it. I say d don't talk to me when you got those things on. But hey, yeah. I'm just an American.
Sophie: Yeah. Yeah. And I think yeah.
Damon: Ha ha ha ha ha ha ha ha ha
Sophie: Yeah. I and but I think more more people will do. Like in in England, we are terribly British and we would rather just complain about it for the next week than say something. Like we we aren't so good at at those more direct interactions.
Gene Feldman: But but to be fair, quite a few but to be fair, quite a few of you have become very Americanized over the World Cup, which has been fantastic to see.
Damon: was rude yanking someone.
Sophie: In England we have always loved soccer. It's in our veins.
Gene Feldman: no. Well I'm t I'm talking about the ranch dressing epidemic.
Damon: Hahaha!
Mike Vann: Yeah, and the Buckeys.
Sophie: yeah. Yeah.
Damon: my goodness.
Gene Feldman: I'm watching I mean, the TSA had to come out and make an announcement about rantressing was hysterical.
Sophie: Well, I've seen when I I was in America last week and the week before and I saw some of the like the news segments about it. and somebody mentioned Buckeys and I've had a Bucky's before. I love Buckeys. The best burger I've ever had was at Bucky's, somewhere in Texas. I couldn't tell you where I was on a massive highway somewhere with massive trucks. But yeah, so I am a I'm Yeah.
Damon: describes Texas to the T right there. mean, you know.
Gene Feldman: Absolutely. those people don't realize the truck stops sometimes have the best food on the road.
Sophie: Yeah, Buckies. Yeah. But I mean, going back to your to your point, Mike, I think and I'd also be interested to hear all of your opinions on this as well. That if I think from like an enterprise point of view of them buying the glasses, I think there's certainly less friction than say like an XR headset. It's definitely a more natural next step than going from like viewing 2D content to being
Damon: And the Buc-ee's is not even a truck stop.
Gene Feldman: No.
Sophie: In 3D content. Like this is definitely like a natural next step. And I see a wider range of day-to-day use cases. But even from like an a company privacy point of view, if suddenly lots of your team members are have these glasses and they use them, what does that mean for your confidential company data and your clients' data? So then will enterprises start giving people like work AI glasses. But then you have the opposite where the individual team members might think that you're spying on them through the glasses. And I think there's still a a trust element to overcome or or even if it doesn't if AI glasses don't record. So you can still get things in front of your screen but doesn't record or it just records audio. I think there still needs to be some hardware adjustments, in my opinion.
Mike Vann: Absolutely. I think we all have thoughts about this. I know Lucky you're into some of this. I mean I th some of what we're hearing and seeing is that that's exactly where it's going. The ability to have visuals, but no recording and ins and no recording of voice or video. but the ability to have to have well, to have audio and visuals, but the power of what you're doing is the ability to have conversation and learn. I'm just not sure
Damon: No, no, that's right.
Mike Vann: They want people out in the public having these conversations, right? So it's interesting.
Sophie: Mm-hmm. Yeah. And ultimately, I think, say, in terms of what we do, ultimately the glasses are there to augment reality. So in terms of what we do, if virtual speech was was to go on like AI classes, it wouldn't have to see the conversation live. You could just ask it to to help help you prepare for it and then speak to it afterwards like a coach. It doesn't need to, it doesn't necessarily need to be in the room. And you would just be in the room in the online version or the XR version. so yeah.
Dan Meeks: Well, I I got a crazy idea for ya. D t tell me if this is just outlandish. But what if you did use one of those glasses as a surrogate? So I I know that you have a virtual environment where you practice the speech and interaction, but what if you have a role player that is getting their prompts inside the display of what to say and how to say it so that they actually get that human feedback, no kidding feedback, but it's an actor that's that's doing it instead. What what are your thoughts on on that? That's kind of a
Ron: question for you.
Gene Feldman: You're crazy.
Dan Meeks: Crazy idea. If you saw the movie Sarrogates, it's kinda like that kind of idea.
Damon: Hmm.
Sophie: Yeah, so yeah, they wouldn't actually be practicing with yeah, I guess it's like AR really, isn't it? But with the but through the glasses, yeah.
Dan Meeks: Yeah. So instead of the the virtual person that you're talking to, there's an actual person and they're getting prompts from the feedback that the classes are listening to. And they're providing but my wife's a nurse educator and they they hire role players to be patients, for instance. So to to have AR powered classes that are saying, Hey, respond like this warmly, you know, and say this exact phrase and they just do it. You know, what are your thoughts on that?
Sophie: Mm-hmm. Mm. Yeah, that's a great idea.
Dan Meeks: Okay. Go go make millions. Go, go.
Gene Feldman: Ha ha ha.
Damon: Hmm
Lucky Gobindram: Well
Sophie: Hmm.
Mike Vann: It it it's all about
Gene Feldman: Now any ideas
Mike Vann: privacy
Gene Feldman: that come
Mike Vann: and legal and risk. It's at these companies, right? Like it it always comes down to that. But but the devices are are out there, right? Like the Evan Realities has the display with no cameras. It's really good. you could build you could build custom applications like this. So it's it's all possible.
Damon: You
Sophie: Yeah.
Lucky Gobindram: The the Evan realities are sweet for Mike, like you know, but it's it's
Gene Feldman: Yeah. And they even and and by and I think where I see you're going too is 'cause like they have like the new James Bond video game has these has the glasses integrated into the gameplay that having that additional layer to where you could say, you know, like going into Riverside, do you have headphones? Do you not have headphones? You know, going to virtual speech, do you have Evan glasses? Do you not? And they could actually have so they could practice with the interface. But But that this actually is going to lead into a next question, though, that's talking about getting into bigger companies, working with all these. How do you stay so small? How how well as somebody that's getting started and you, you know, you've really kind of kept a very independent role in this. So when you're like reaching out, starting to think about okay, do I which glasses should I reach out to? Which partnerships should I go? How do I avoid the big guy? coming in on me. How do you do it?
Sophie: So for for context, we so there's twenty-one of us on the team. So we're still a small company, we are entirely bootstrapped, so we've never raised any funding and we were actually very intentional about that as well. I think what's the expression around like you put something in the deep end and you either learn how to to sink or swim. And I think if you if you don't have VC money, you learn to swim. And if you do, you like
Gene Feldman: Yeah.
Sophie: You might doggy paddle, don't know if that translates in America, you might doggy paddle, you might tread water, and for some you then learn to swim, but for others you need a life raft to keep you going. And then each time you paddle, you might need another life raft, and you're just constantly getting to the next life raft, which is like money from VC. Some get there, but most die trying. And I think by not raising VC money, it has meant that we have learnt to swim. And we have always gone in the direction of customers because we don't have a fallback, we don't have fluffy metrics that mean nothing when you haven't got money. Like you don't go broke with money in the bank account. And I think by building based on what our customers need and not getting distracted because we've got a cushion behind us of ten million dollars. Like it's very easy to get distracted and to get frivolous. And actually, if what we found is if you Follow what the customers want, the customers will stay, they will refer you, and and more customers will come with that. And ultimately, it's I guess it's being like disciplined with ourselves that we need to work on something. We need if it doesn't bring us revenue, we're not gonna work on it. So if one customer out of 178 has suggested a feature and they're worth Like are they a low value customer? We haven't got time like we have a small team, we haven't got time to create that for them. so yeah, I think it's about ultimately following the direction and the feedback of customers, moving quickly where we can, and not getting distracted by shiny objects.
Ron: So question on that, Sophie. So that's a very prudent strategy to do what customers ask. But at some point, you're getting a lot of feedback from this and you're doing your own thinking. And at some point you're doing a look ahead and like having a vision about where you want the company and the products and the capabilities and the value being delivered. And so you have to balance off you keeping it, running it tight or lean. But I got this vision. So in three, five years or so, I could break this market open and buy this capability. But I got to invest. So how many guys dealt with that?
Sophie: I mean our vision is focused on impact. So I mean our like previous vision was to help over a million learners worldwide. As we're approaching that, that number is shifting. but I think yeah, I think it's it's having the vision in mind and I mean I mentioned right at the start about always being like ahead of the curb and innovative and we've been the first to market in many respects, which has helped us position there. We do have to make calls of what we think might be best. So say for example, like AI, like Gases is an example of not a single customer is asking us for that, but with more knowledge than the most of our customers have of say like the XR space, is that the next step? Is that a leap of faith that we have to take and a educated judgment that we have to take and ultimately hope they hope they follow at some point. but also I think our business model, we we shifted our business model four years ago. So it was it was still just me and Dom up until four years ago. And then we shifted our business model to be recurring revenue. So we always have a very sound knowledge of like the financial side of things. We do we obviously have like built in a buffer. We very much have an ethos of when we hire, we will never have to make someone redundant. Like I think it's very careless when companies do that. I understand why companies do that, but that's not something that me and Dom ever want to have to do. Like if we are we have twenty-one people and we are protecting those twenty one people and every decision we make is to protect those people and therefore the the product has to be this way, the customers have to be serviced this way to to naturally flow into that. And with that we have like we have a buffer. We haven't raised VC funding. But for example, our largest client, we have not upped our expenditure or team to reach that limit because if we lose that largest client, then we're in a bit of trouble there. so I think Where the industry has been the last like ten years when we've been in it is quite different to how it how it probably is going to be over the next couple of years. So I mentioned earlier on that our competition isn't other XR companies, it's there's quite a few now like AI conversation companies cropping up. And it's it's become easier to make, say, AI role plays, but it's still just as difficult to make good ones. And our moat is the quality of what we do. We have a learning design team. We're not just saying like two technical people who think we can make a quick buck by doing this. We are a team of people with different expertise. And a crucial part of that is the learning design team that are bringing the actual, like tangible, like education side to the app.
Ron: So extending on that, where do you see the major opportunities for your company and for the immersive tech space of AI market in general?
Damon: say two years, 20 years, put on your wizarding hat and look into the future. Use your imagination.
Ron: Yeah, have a seat.
Gene Feldman: Ha ha
Mike Vann: Twenty years.
Ron: Yeah.
Sophie: think something that is different now, even t compared to last year, is I don't think you can plan for five years, ten years, twenty years, because everything is moving so quickly and in ways that we can't imagine. So in you know, I joked, well, like I said earlier about you can't tell like one in seven billion people's voice just like know who it is just from a voice. Like
Damon: go. Yep.
Sophie: Well, yeah, you probably can in a couple of years, or as one of you said, maybe in a couple of months. so I think it's very difficult now to plan further than like two years away. So I mean, what we do is we like focus on our vision, so reaching and and helping as many people as possible, but also focus on our ICP. So what's important to them. And say we work with enterprise in higher education, but ultimately our ICP is enterprise. And how can we make our product more enterprise friendly. So is that like ISO and SOC two certification? and again like building that moat and trust and that reputation as being an enterprise ready platform, not not one of these companies that's that's popping up quite frequently now. I think I said to Jean when we were like preparing for the podcast that the the newer companies that are popping up obviously there are some good ones, but there's more not good ones. And it doesn't worry me that our customers or potential customers might choose a competitor because competition is just part of life. what worries me is that they will try one of these more Mickey Mouse pop-up ones and similar to people ten years ago who tried a VR roller coaster and then didn't touch VR because it makes them feel sick. Like my concern is more that they will try one that isn't good and get a bad taste in their mouth and then think none of us are good. so I think it is a very interesting time, and I guess to go back to your question, like it's very hard to predict.
Gene Feldman: Yeah.
Sophie: what is going to happen, but I guess if we focus on our like our strategy, rightly or wrongly, is to focus on our North Star of our vision and our customers and just keep going from there.
Gene Feldman: No, I I Sorry, Ron, I cut you off. Go ahead.
Ron: Yeah, sure. So Sophie, who do you engage with in enterprises?
Sophie: learning and development teams.
Ron: Do they have much clout?
Sophie: Yeah.
Damon: Can we answer that for you? No. Don't put her in that position, Geez.
Gene Feldman: I I I was yeah, as
Sophie: Che?
Lucky Gobindram: Ha ha ha ha.
Mike Vann: Ron Ron's grilling her on her clients. Come on, Ron.
Gene Feldman: Sophie, I'll take Sophie as the as the as the learning and development expert, I'll take this one for you. Okay. We have influence and that's about where it stops. We're told to find the really good ideas, and it's our job to find like people like Sophie. And I've brought I've brought her into some of my past roles. And it's and then she t has those
Mike Vann: Don't answer, Sophie.
Sophie: Ha ha ha
Damon: No.
Gene Feldman: engages in those further conversations, but it's all based on really we all we have is really influence. But if a d if senior leadership has had a bad taste in their mouth from it, or if they read some of the articles like on c on CERN other integrated device a AI systems that you can kind of mock a simulation together using it, they'll start thinking it can be homegrown. So it's it's really us to kind of position Sophie as that expert for that type of solution and why we don't want to use those other ones.
Damon: Well said, well said.
Sophie: Yeah, and I think
Ron: So.
Sophie: it comes back to the the customer that I mentioned with the the feedback from the distribution centre that it's it's going to become increasingly important for virtual speech and and similar companies to be able to help that customer success element and to be able to help them with that implementation and help them prove ROI. And I mean, I spoke about like percent improvement in skills, percent improvement in confidence, but ultimately it needs to be tied to money. Like that's the most important thing to to a lot of enterprise, like people who hold hold the budget. and I think in general, I I was at a talk a few months ago and I think it was from like the Institute of AI or or something. somebody very high up in the AI world and they were saying that This year is is the year of like the reckoning of AI. Like people have tried these things and now they need to see ROI. And I wouldn't call us an AI company at all. We are a learning company. We happen to use XR. We happen to use AI because those are the best tools to achieve what we and our clients are trying to achieve. But we ultimately also need to be able to prove that value and in as many ways as possible because talking about skills and confidence will appeal to one person, but that's not appealing to the CFO. They care about how much money is is being saved or spent.
Damon: Yeah. Yep. Yep.
Ron: Well, the CFO cares about cost, they care about ROI, they care about managing risk. So if I have a better learning environment, they learn faster and learn better, that enables the corporation to better understand and manage risk. So actually you're in the business of meaningfully improving outcomes in enterprises.
Sophie: Mm-hmm.
Ron: And through the learning and better managing risk, you have better outcomes. And that's kind of how the executives want to hear about this and the ROI thing, you know, like new opportunities, new value, things like that. And one of the other things about this is even your own company, as well as enterprises is like people learning is an interesting thing. I've observed people don't ask near enough questions in their own best interest to learn near fast enough or develop these streams of consciousness to really get into a domain and learn quickly. So this is kind of paving the way for these AI agents to come into place. And unless you learn fast and have real insights, many folks are at risk of either being augmented by AI agents or replaced by them. So it's kind of a bit of a, you know, not pleasant conversation. But if you were running a Fortune 500 company making billions of dollars, you're looking for insights and ways to make the next billion bucks in revenue and 100 million on the bottom line. And if the people can do it, then great. If with technology they can do it, that's great. Or the technology of increasing larger role on this. And a big part of this will be determined by how fast people learn and how well they learn and develop these insights about where the money is going to be. So do you have any insight about how you contribute to that?
Damon: next hour.
Ron: Pfft!
Gene Feldman: Yeah.
Lucky Gobindram: Ha ha ha.
Sophie: maybe to part two. yeah. I mean I Yeah, I think that's a really good point. And I think it is those skills around like speed to learn, like mindset, like growth mindset, adaptability, which are all essentially those soft skills. something that I think can can help in terms of those efficiencies is
Ron: You
Gene Feldman: I know, we we may have to have you back, buddy.
Sophie: We like we have some clients, for example, who will directly tie to like staff turnover and like how much that costs the company compared to like say when they use virtual speech, the cost of virtual speech, or in the oil and gas sector, it will be or it can be the like huge percentage of the like learning costs is actually spent on the logistics of getting people to. the learning centers because they have to fly them in from all over the world. It's technically when they're off duty so they then have to pay them pay them for those days, the accommodation and everything. And one of our clients told us that 70% of their cost center is actually spent on the logistics of the learning. And so we are coming out of that cost. So if say, I mean it will be more than this, but say it just for easy numbers, like they have a 10 million budget for Training, if seven million of that is going on logistics, then paying seventy thousand dollars for virtual speech is a no brainer, plus the headsets obviously, but like it's suddenly it's it's a no brainer then. so I'm not sure if that does answer your question.
Damon: Well, it's going to lead to a part two. I'm sorry. We just can't teach. That's an MBA discussion.
Gene Feldman: Yeah. it
Ron: You
Sophie: Yeah. Maybe we'll park that one. We'll come back.
Gene Feldman: No, it a and great conversation. Unfortunately, yeah, we it we've to go longer than an hour, it's just like I know we're gonna kill people. but no cause we get carried away. So yeah, so we're kind of at time. So I'm we're gonna call it here. But Sophie, we would love to have you as you can tell, we would absolutely love to have you back. If you ever wanna come back or if you actually or if there's information that you wanna share with us, we like I said
Sophie: Ha ha
Gene Feldman: the chairs always open and we'll talk after the call. but one of the things that we do like to ask, one of the things we like to ask our podcast guests is for a final protocol that based on, you know, you've had a really interesting story and we've kind of hit you with a barrage of questions. But if you were sitting across from you when you were 22, what advice would you have given you?
Sophie: Thank you very much.
Damon: That's the final protocol, All right, all right.
Gene Feldman: That is, I like that one.
Sophie: Oof. What five?
Gene Feldman: I like ca just because Sophie, I mean, she hasn't had anybody in front of her doing any of this. So I'm really curious that what advice she would give herself looking back.
Sophie: I mean I guess so I actually have this is very hard doing I it's not necessarily it Yeah Yeah I think I would I'm like toying between a couple of things. So I think one thing I would tell myself is to keep going, which obviously I have done, but I
Gene Feldman: I'm I'm watching you too. I I wanna throw you a lifeline but I don't but I'm having fun.
Damon: I'm not sure that was a compliment, Gene.
Gene Feldman: I know. Please.
Sophie: I wouldn't tell myself how hard it's going to be because she wouldn't have done it. And like and I've enjoyed and I continue to enjoy it and I would do it all again. Like I love to learn and I have learned so much. Like I this was my first job. So when our first customer asked for an invoice, I literally Googled what to put on an invoice and I was using like a website, yeah.
Gene Feldman: Okay, fair.
Sophie: Like I because I'd never seen one before. I was at university. Like and and like so I mean that that tells you from where I've started. and also that also to be honest, just highlights we need to learn that sort of thing at school and university. I was doing a master's in international business and no one had ever shown me an invoice. Is wild as well. so obviously I've learned a lot along the way, which I've loved, but
Gene Feldman: Yeah. Absolutely, said
Damon: That's true. Mmm.
Sophie: There have been obviously like I would say like startup life, especially in the first few years, is very lonely. And so yeah, I think I the advice I would give to myself is to keep going. And then I also have a couple of quotes on my notice board, which have literally been there since 2017. And they are around the sense of like you grow the most in the discomfort. So, and you know, everything you want is on the other side of fear, like those kind of things you might see on someone's wall. but I do totally believe them that and if I take it back to how I started and confidence, like I had a fear of public speaking. I had a fear of speaking to and public speaking isn't just a presentation or, you know, a conference. It's like things like this. And I was scared of that. But everything I have achieved since then is because I'm not scared anymore. so I would Tell myself then and probably also tell myself now these things to just remember.
Gene Feldman: I love it. Absolutely love it. So for those that are listening, we this we've spent the past hour and a little bit extra with Sophie Thompson. So please feel free, give a listen to the channel. Please follow. Share. It's gonna help us grow. We're a small little tech podcast, but we're trying to come at this from a whole different angle. so please share and support.
Damon: There you go.
Mike Vann: Very good.
Gene Feldman: if you have any questions or if there's anybody you would like us to reach out to or if you have a story you would like to share, go ahead and reach out to us. go ahead and leave a comment on any any of the channels that you watch us on or go to our LinkedIn page. But Sophie, we thank you and to Neon Protocol team, we absolutely thank you guys for joining us today. And let's have a great week.