Steven Webster - When Fitness Products Start Acting Like Coaches

All right.

Let's dive into it. Ravid fire questions. You go first, Marco.

Yeah. The first one is I don't know how to pronounce this in English because um in in Spanish you say karate, but maybe in English karate? Karate. But it's the pronunciation's good. Karate or jiu-jitsu for you.

You know, when you're a white belt, they look very different. When you're a black belt, they look exactly the same. And so uh yeah, I have an instructor that I used to train with Patrick McCarthy, and I'll I'm gonna steal his line to answer the question. A punch is a punch, a kick is a kick, a joint only bends one way, and pain doesn't discriminate. So uh I'm not choosing.

Love it. Uh here comes the next one. The US or the UK.

You can't do that. You can you just can't do that. I think you know the Scottish people are the funniest, friendliest, most fiercely loyal people. Edinburgh, whereas you know I grew up outside, but spent a lot of time living in Edinburgh. I think it's one of the most beautiful cities in the world. California is like one of the most beautiful corners of the world, and I'm a techie that lives in San Francisco, Silicon Valley Bay Area. So I love being here. So uh again, you're not gonna make me choose. I just feel very lucky that my journey has taken me from one to the other.

Are you still doing a lot of back and forth or maybe? Not so much lately.

Come on, you know what you know what it's like having a startup salary. You don't travel as much as when you have an Adobe salary.

I agree, I agree. All right, the next one. Microsoft, Adobe, or your own startups? Best years.

Not Microsoft.

Didn't you say before it's like having Microsoft on the resume on the CV?

Yeah, listen, I you know, I never imagined I would have gone to Microsoft. That's another story for another time. But you know, I followed someone that was consulting uh Adobe. And when even when I took the interview, I was like, you know, you've heard about the Microsoft interviews. It's like, you know, it'd be good to kind of say that I did that Microsoft interview that, you know, uh, so I flew up to Seattle and did the interview, and then there was a moment in the interview there was one individual who interviewed me. I was like, wait a minute, like this would be huge to kind of try and do this. So, you know, I've got some great friends and I had some great times at Microsoft. But to answer your question, like Adobe just changed who I am. It was just such a fantastic experience, fantastic company. I think you guys know, you know, they acquired my company. So we were like a small company that kind of came into macro media, and then immediately Adobe. We knew that Adobe were acquiring macro media. And I was just really lucky to have a bunch of like leaders who were kind of this was probably going to be the last job in their career, just take me under their wing. Like I was the I was the rookie on the management team, on the leadership team. People really kind of invested in me and kind of coached me and mentored me. And I couldn't have done Asensi without all that coaching and mentoring and learning that I did at Adobe. But listen, you know, Asensi, let's be honest, it's you know, when you quit a job at Microsoft because you believe something's possible, that like ask a hundred people and they'll tell you it's it can't be done or it shouldn't be done or it won't be done. You know, I've got a tremendous amount of pride that we've got this far. We've done exactly what we said we would do. Uh anyone that met us 10, 12 years ago, you know, they look at us now as like you guys actually went and did it, and you've got great customers, and the technology's real. And so it's hard not to be incredibly proud of that. So yeah. Anything beginning with A is the answer.

That's a great.

So they think about the next companies, what's it then? Uh anthropic. Right. I want to share.

Anthropic.

First one that came to mind mine was Adidas from the from the sports world. But yeah. Awesome. Uh changing gears a bit to the fitness-based fitness world, most overhyped fitness trend you see going on currently.

Scores. Everything being a score. Like it's just like, you know, we're diminishing the data, you know, and and scoring things that, you know, in the past might have been the something that like an elite athlete would care about. Like it's a marginal gain that matters to them. But I think we're just obsessed with scores and not obsessed enough with, you know, I think about syllabus rather than scores. I'm much more interested in whatever that thing is somebody's decided they want to do, lay out a path for them to get better at it. You know, as a martial artist, I think white belt to black belt. Everything for me is like white belt to black belt. Like, because the nice thing about that is, you know, you linger as a white belt for a long time. You know, you know where the bar is, but you don't worry about all the other stuff until you reach the bar. Now you're a yellow belt and you just keep doing that. So it's these, you know, rungs in a ladder, if you like. So I would love to see us move away from this idea that everything's a score, and like, you know, Marco, what's your sleep score? Let's compare your sleep score. I sleep better than you. Like, what does that even mean? And related is can we move away from dashboards? I like to talk now about move from dashboards to chalkboards, right? You know, or or to or to whiteboards. Like, you know, I mean, I coached you know, decades of martial arts. I coached at the University of Edinburgh. We won 10 consecutive national championships. I never once made the students stand at the wall and look at a bunch of graphs of like the speed of their punch or the height of their kick or the you know flexibility range of motion when they're trying to do the splits. You know, we would just have coaching conversations. You know, we would observe things. It's like, hey, I want to show you something, and then we would move on. And so I think, especially in the world of we're moving away from apps and we're moving towards agents. Uh, you know, I have a strong point of view that I've held for the last 12 years, if not more, that interaction design will move away from apps to things that are conversational and gesture-based. We'll speak and point to the machines in front of us. We won't like like type and swipe with them. And so when you do that, I don't need to be staring at a screen and a dashboard. I don't need you to pre-imagine all the different things you might want to explain to me. AI is more than capable of saying, let me draw that for you, or or let me kind of just show you that quickly in a little chart and then get rid of it. Wipe it off the chalkboard and wipe it off the whiteboard. So uh yeah, that's my provocative point of view. I don't want to see Ura and whoop dashboards, like just speak to me like a human, speak to me like a coach.

That's great. That leads me to the next question, which is like a coach or machine who gives the better faith back today.

Yeah, it's it's a nuanced question, right? I am quite confident my my tunnel through the wall does a better job than a human in the moment, in that rep of deciding whether to keep the weight where it should be or to start to drop the weight. Yeah, that's something a machine is just very good at doing. And I'm also quite confident that, you know, Marco, if you wanted to teach me some jujitsu on the mat, I'm quite confident you would do that better than a machine would do right now. Let me kind of show you how to slip a guard. So I think it really depends on the context. But what I will say is in every industry, including ours, everybody wants to believe that they're the only industry where the labor market's not going to be touched by AI. Oh, it can't do our job, right? Everybody says that. But it's taking jobs one after another, or rather than taking jobs, because that's not my point of view, it's stepping into the workforce alongside human workers at an incredible rate of knots. And so I think the coach versus machine, I think really the way I would answer that is machines are only getting better and better and you know, being imbued with more and more ability to be able to do the job incredibly well. And even humans, right? Let's be honest, like humans are on a standard. You know, one physical therapist might not be very good, and another might be incredible, and then you you refer them to everybody. So even if a machine can just meet the bar of the lowest quality of human, that's a pretty impressive bar to reach. So again, I'm not giving you an answer to any of your questions. They're all like No.

You're doing circle answers.

Finding elegant ways, the political way to escape.

Yeah, but that's that's a very good interesting because nowadays it's like I think AI and everything will help, and the machines, of course, will help you. It's a tool at the end of the day. And a very good tool because they have a lot of information. Now you as a human can take a better decision with everything. So it's like uh it's uh it's gonna be a symbiosis between both things.

I mean we've been here like forever. You know, it was calculators, and then it was computers, and then it was the internet, and then you know, we we always have this portion. You know, everybody adopts technology at a different rate. The rate of diffusion of technology is different, but it would be um it would be strange to imagine that that AI is not gonna deliver on its promise.

It's gonna be a very interesting future. Oh yeah, for sure, for sure.

You mentioned something interesting there, Stephen, when you said like okay, you believe like in the future we're not gonna swipe and touch anymore and like give instructions to machines or the kind of surrounding world and environment we're we're interacting with. But it's gonna be more through through voice. Like what's so you're thinking of like Alexa Siri type of like you talk to the machine, you give instructions, or even more conversational where it's a back and forth, or like what's what's what what do you envision? And then maybe boiling this down as well to the fitness, wellness world that we all play in.

No, I mean for sure. Listen, will the keyboard go away completely? No. I mean, sometimes like you know, for me I love to write, and like typing helps me kind of slow down my writing to a bandwidth. I've tried to dictate my writing and it just doesn't work because as soon as I start speaking, you know, I'm communicating kind of differently. So I'm not saying keyboards will go away forever, but I would say I spend the majority of my day, and it this is a muscle that you have to choose to build, right? This is not something that you like just one day you do it. But I use Whisper Flow just now as an app that I speak more than I type now. You know, if I'm writing a document or if I'm creating a presentation, like I speak to my machine like I would speak to a human. Uh it's like I'm kind of thinking about this and and I'll correct myself mid-sentence, and uh, you know, and so I'm just constantly in conversation and dialogue. And then you to answer your question about kind of fitness, I I always come from what's the best human-to-human experience and how do you deliver that human-to-human experience digitally? And again, any of us that have ever trained in any sport, it was never like, okay, go back to your screen and look at the screen and watch the video, and when you're ready, go and give that a try. Like that's that's how we had to do it with the technology we had at the time. But if you know what you're doing and you just want to go through a workout, or if you've been shown your physical therapy and you just want to go through your program, a coach in your ear is the most natural way of having that experience. It's like having a trainer, you know, sitting in the room with you. And then the most natural thing to do if you're stuck is not go to the screen and like where's the help icon and where do I go in the hamburger menu to find da-da-da. Like, that's not how we think. It's like, hey coach, could you show me that again? Or am I doing this right? Is it this foot back or this foot forward? Like, we just that's the natural way. And it's kind of interesting, like, we'll need to relearn natural communication because machines have forced us to meet them where they are, but now they can meet us where we are. I love to think about like the pinch to zoom on Apple. That was a great example of for many of us, that was by no means a natural interaction. Like we had to be taught slowly how to use a touch. You can touch, oh, you can swipe. Wait a minute, you can use two fingers, you can do what? And if I can do that in a photo, does that mean I can do on a on a video? No way, I can do it on a video as well. Fast forward to today, we've all seen a kid walk up to a TV and do this. Or I remember my son on a magazine doing it on a magazine. Like it was so intuitive to him, like trained on an iPad. The judgment, I can feel it coming through the screen. But uh, you know, he was so used to what he could do on an iPad, it made no sense to him when he was like, you know, two or three years old that he couldn't do it on a magazine as well. Like, why can't I zoom in on the picture? So I really think about you know, that we're gonna have to relearn how we interact with machines, but in the process of that relearning, the interaction modes feel much more fluent and much more natural and frankly much more enjoyable. I love sitting down at my machine and just going back and forth and brainstorming and ideating and like, all right, could you capture that and share it with Ross uh as a PDF? And it's like done. I didn't touch the keyboard. I have a foot pedal down here. I can press the foot pedal to talk. I can sit back in my chair and talk.

That's interesting. It's a new way. I mean definitely I think everyone is preferred to speak rather than you know, type or write also. But I think also the this the brain needs to start using those type of things. I don't know. It's going to be very interesting for the new generations. I can see a lot of my kids talking with Alexa in a very easy way, or with ChatGPT talking about so many things, and they are not looking for in the iPad, but because they have iPad, but at the same time they have Alexa and they have uh open AI and ChatGPT, all these type of things. So now it's more conversational. It's really interesting right now.

Yeah, no, it's we're we're we're living in crazy fun times.

It actually reminds me of a situation I had this morning in the in the gym, and I'm not making it up because I'm trying a new program routine, and there was one work out. I don't I don't recall now which exercise it was, but it was an exercise I I didn't recognize. I was like, okay, what's that? And first thing I did, it was just like the voice recognition like recording for for ChatGPT of like, hey, what's that workout? Like, give me an example, show me the movement pattern. That's and that's still some you know steps to go. Like if you could have it obviously even more seamlessly somehow incorporated, of like, okay, I should be now you know doing this workout, help me remind me what was the workout like, what's the movement pattern? Oh, go to this machine, oh move this, you know.

A big thing with AI, right, is you start to use these tools more, and you you know, I think a big unlock is when you move out of chat and you move into co-working, whether that's codecs or co-work, when you move into this idea of like, you know, I'm not chatting and copying and pasting things in and out to a to a chat assistant. We're actually in it together, we're in this environment together, we've got shared context to use the term. Well, think about when you're in the gym, and now when you kind of pull up that moment of confusion, like I don't know what I'm doing, you shouldn't have to explain that I'm I'm on a new program and there's this exercise and I don't exactly know what the exercise is. Could you show it to me? All that context should be there in that moment because you know, the agent, the AI coach in your ear knows where you are, knows where you are in the gym. You had Brian from Lifetime on your show, and you know, he's thinking about like, do I know which piece of equipment you're at right now? So I know which piece of equipment you're at, that's context. I know your workout program, that's context. You're telling me you're stuck. That's all I need to know to say, like, oh, you're at the rowing machine. Okay, just sit down, let's go through this together. You know, get to the front, grab the handle. This is called the catch, right? And then we're straight into a coaching experience. So uh that's one of the wonderful things about AI is the more context we can have around the moment that the person is in, the more natural the interaction can be.

That's awesome. How's this kind of tied to a sensei? Maybe you can give us just like the 30-second pitch on what you guys are up to, how this is all connected, and yeah, what you're bringing to fitness, wellness, health.

Yeah, so for anyone that doesn't know us, um what? No, for anyone that doesn't know us, we've invented speech recognition, but for movement, we do it with a camera instead of a microphone. And we do it so that brands in health and fitness can create AI caregivers or AI coaches that understand and can intervene and can assist in coaching us through movement. And so, you know, we started in fitness and sport, and we today, you know, we work across sport, fitness, healthcare, healthcare being physical therapy, like musculoskeletal, but also neurological. We've got customers doing stroke rehab and you know, helping Parkinson's patients with motor control and slowing onset. And so throughout our lives, we need assistance from another human being to show us how to move a certain way, whether you're learning to dance at your wedding or you're strength training for the first time, or you're recovering from a surgery. And so we just think that's a labor market, that there's a a growing labor gap, and it's a real opportunity to fill that labor gap with agentic coaches or AI coaches that live alongside the human workers that are churning out and burning out today.

I really like it, and I do believe that you are solving a big problem. You know that pr just recently I tore my NCL, the ACL, the meniscus doing jujitsu and everything, you know. So after the surgery, I was in the process of a lot of PT. And use this type of technology that also can help me to improve everything in my own home. It's a big difference. I mean, you need it because you can do all these exercises in your PT, but also in your home, you have to start doing everything. And you need someone who can tell you if the movement is in the correct way or not. So I do believe you are solving a very big issue.

Yeah, because a huge problem with PT is people just don't do it. You know, the adherence problem, right? And it's not that we don't want to get better. Like the the most one of the most common things you hear people say about why they didn't do their PT is I didn't want to make it worse. Like I didn't really quite know what I was doing, and I didn't want to so it's just easy to cognitively offload that decision to like, well, I'm gonna see the PT on Tuesday, so I'll just wait till then and I'll do it then. And so half-assed, right?

Like not the right amount of effort, not the right amount of like weight, even because they I wasn't sure. And then it's easy to drop, you know, I'm doing it without weight because it's lesser like strain on the case.

No, you're not getting the adaptation you're looking for because you're not using the right weight or you're not doing the right reps and and stuff.

And then you cannot finish the PT because the insurance policy will not pay everything. So it's like uh I mean, it's like too many things at the same time. It's a perfect storm, so you cannot continue doing and improving yourself. So that's why this type of technology will really help other people just to start building themselves.

Exactly. Exactly.

But help me to understand how do I connect the dots. I mean from Microsoft, well, karate, and then Adobe, and then acquisition from Adobe, and then a sensei. How does that happen?

How does it happen? Yeah, I I mean I think there's a common theme coming into this interview. You kind of made me like think hard about a few things. And there is a common theme. And let's listen, I lean on my martial arts quite heavily, just you know, a lot of my philosophical approach to how do you do hard things. I mean, you do jujitsu, right? Like that's a hard thing, and you know, you you you have to have a kind of a mindset to approach doing that as much as maybe a physical exercise. And so there's a con there's a concept in martial arts called shuhari. It's three words, shu hari. And shu is the idea that first you learn from the tradition. So, you know, you're a white belt, you're just learning like this is how you do the basics. You learn the tradition. Ha is when you start to break from the tradition. Uh that might be where your own flair, your own finesse, your own skill, that's where we start to look different from each other. That person kind of you know is different than the other person. And then re is where you transcend the tradition, you become the transition, the way that you do things becomes the new normal, right? And it's very, very kind of circular. And that's how I approach martial arts. Like when I coach martial arts, you know, I learned jujitsu, I just did it the way I was told to do it. I did the same with karate. But then there was a moment when I was coaching of like, wait a minute, I've got my own take on how you should teach this stuff, both in self-defense, which is very kind of practical martial arts, and in sport, which is like let's go win all the medals. And in both domains, I had kind of very I had the courage to have a provocative point of view and follow through on it, right? And I learned that on the mat and it showed up in my career as well. So Adobe and the way I ended up at you know, the company I started was essentially a provocation, you know. It was a I don't think you know. I'm about to say something that you know anyone like younger than me is probably gonna be like, well, that's obvious. But it wasn't obvious at the time. You know, my provocation was like software is not gonna ship on a CD that you feed into your PC and install, right? Everything will be delivered through a browser. And at the time, browser sucked. Yeah, it was dial-up modem. Every time you clicked a button, it would go to the server, and 30 seconds later the page would re-render. You know, I used to describe it like reading a book through a straw, you know, you would just like you had no idea where you were or where you were going. But I had this provocation that you know we should be able to deliver really rich desktop-like software in a browser. And when I was describing to people, like, what do you mean, like Microsoft Word? You're gonna have Microsoft Word in a browser. I'm like, yeah, absolutely. There's no reason why not. You know, this is in like 2002, 2003, 2004. And now we just accept that you can run Photoshop and you can run Word and you can run Figma in a browser. But that was a really provocative idea at the time. And that's how we got acquired by Adobe, because I was using Adobe technology. I was using Adobe Flash as a front end to enterprise software back ends, right? Which was just like who would put Flash and Java together? That was kind of not conventional wisdom. And then fast forward to you know all the way through to Asensi, it started with a provocation. Like, I believe that you'll be able to learn, I believe that you can democratize access to world-class coaches, that you can take the knowledge and the know-how of a coach and deliver it as effectively, digitally, to everybody, everywhere, all at once. And even today, there's plenty of people who want to debate and discuss that. But uh, we believe it to be true and we've delivered on it so far. That's the through line.

That's amazing. Um it's been what, 11 years already, 12 years?

12, 2014, 2014. Although the early years were a lot of can this even be done. You know, there was definitely kind of two or three years of you know, just the c the the three co founders like is it is there even a business here? Like is it Even technically possible before we uh 12 years old. It's a journey.

It's a big journey.

I know. I know. I look forward to our 20th birthday.

Yeah, well, help us just understand like how do the experiences really come together? Because obviously you're operating mainly today in a B2B fashion, which I believe also hasn't always been the case, right? So maybe you can give us a bit of the backstory to it. And now in the B2B setting, how do you how do you bring those experiences to life that you currently can produce with the Assensi Tech?

Yeah, no, so there was a period from kind of 2018 to kind of 2020, 2021. We were still developing underlying technology. And listen, necessity is the mother of invention, right? We were hardcore RD mode, trying to build out like how do we train the way that we can decompose a movement into something that machine can recognize and recognize in a nuanced way. Like it's it's 10 minutes with codex to sit down and say, you know, give me a skeleton and tell me every time somebody squats down and stands up. You can hack a demo together of that, but to have a nuanced understanding of like, can you recognize an Olympic weightlifter versus a stroke patient being asked to do a sit-to-stand? And can you see all the nuances in their movement? Yeah, so when you start to really unpack the problem to be something that you want to solve in a generalized way, it's really, really hard. And so as we were doing that work, we did two things. One, just from a can we bring some revenue in and bootstrap the business, not have to like, you know, VC fund everything. But also, can we work with a constituent, you know, like almost like a microcosm of the broader ecosystem we want to serve? And so we did a very kind of McKinsey-ish Boston consulting group kind of approach. What would be a good niche sport that we could pick that would be learning for us, where we could work with equipment manufacturers and governing bodies and coaches and athletes and really understand what are the consumer experiences that we can deliver and that you know people get excited about. So, long story short, we picked indoor rowing. And at the time, there was no hydro, there was no Peloton Roar, there was no Argata, like there was concept two, and that was it pretty much. And so it was very niche. And we built a B2C app around indoor rowing that was really a way for us to kind of dog food our technology. That's a Microsoft term, but you know, eat our own dog food on the on the technology and just learn. And that app is still in the app store today. There'd be people banging on my door if we were to take that app down. You know, there's there's people that have been training with that app since 2018, and they you know re-subscribe every year because they're they're so attached to the coaching program. So that's the the B2C. But from the day we started the company, it was always our goal to be what I call an ingredient technology. So, like a Dolby Audio or like an Intel inside. It was the only way I felt we could ever reach the scale of my ambition that you know, across physical therapy, occupational therapy, sport, fitness, like we could never be B2C in all of these places all at once. And so we had to be an ingredient technology. And so to answer your question, you know, we're a software development kit, we're a cross-platform software development kit, you know, like yourselves. And people take our technology and they integrate it into their product. We're what's called a headless SDK, so we're not providing a white label experience. We provide all of the real-time information that you need to deliver an interactive personalized experience. And you know, that's a decision that, as it turns out, I'd love to say this was like genius and foresight, but you know, it was like not. Um what we developed was a language and a grammar for posture and movement. So we have a very, you know, kind of JSON-ish language and grammar that allows us to describe any static position that someone is in and the kinematic sequence as people move through these postures. And that's how we chose to solve the movement recognition problem in a way that nobody else does. It kind of helps that my co-founder took a speech recognition company public, so it just so happens that he's thought a lot about like how do you model like stochastic humans in a very deterministic way. So that's the approach that we took. And yesterday, our customers would be taking that real-time data and those insights, and they would power your rep counter on screen or an audio file that's giving form feedback or drive a text-to-speech engine. So that's kind of all of our customers so far, up until I'll call it yesterday. That's how they use our SDK. But having a language is a pretty powerful thing right now because there's these things called large language models. You might have heard of them. You know, LLMs love to consume languages, they love structured language. And so now we just happen to have this massive moat that we didn't plan for, which is we have a language that's proprietary to us that allows an LLM to understand and speak the language of movement. And so today, and you know, we'll maybe unpack this uh uh a little later, but we have an entirely new addition to our SDK, which we call coaching intelligence. And what coaching intelligence allows us to do is allow you to just have these very interactive two-way conversations with a coach, a cloned voice of a coach. It just feels like and sounds like there's a coach in the room that has all that context I spoke about earlier, can see you, knows the sensor data. Maybe it's the sensor data off your flywheel, off your rowing machine, has all your wearable data, has your last year of workouts, knows what program you're doing, and that there's a new exercise in there. So coaching intelligence um ingests that language of movement in real time and just gives you the experience of kind of having a coach in the room. And we're having our pinch zoom moment. There's an opportunity to reimagine what interaction design is like when you're speaking and pointing and recognizing movement. And so I guess I'm giving you guys an announcement here. Didn't mean to, you know, what we're going to be sharing with the industry is here's a set of interaction design patterns that you can, you know, we'll provide them to you. Right now we call it our UX kit. I don't know if that's what we'll call it, but here's our UX kit. It's a set of interaction design patterns that you just instantiate this piece of design. You can style it, skin it, do whatever you like with it. But as soon as you instantiate that design, it just pulls in all of the underlying Assensi technology, the on-device LLM, the on-device memory, the voice cloning, the conversational engine, the movement recognition, the motion capture. Yeah, that just all gets sucked in behind this tiny little component and it just speaks backwards and forwards to you and coaches you. So I'm super excited about how ridiculously easy it's going to become. I mean, you've caught me on a passion subject here, so I'm giving you the five-minute answer. I'm catching myself up. The last thing I'll say is I shared an essay at the start of the year about how I was thinking about like this moment that we're in where everything is, you know, changing all at once. And you know, again, I don't feel this is a particularly provocative thesis, though many might. You know, I think the the dominant coding languages are going to be English, Spanish, and Mandarin. I don't think they're going to be TypeScript and Kotlin and Swift. And so the idea being that under the hood, it's going to be agents that are writing code. So we now think about the Essency SDK as much as like how do we make it easy for an agent, a coding agent, to understand how to use and incorporate our SDK into a product and get out of this mindset of you know, we're trying to write documentation to explain to you guys how to integrate our SDK. We're speaking to your uh have have my agent speak to your agent. So it's a long answer. Does that answer your question? Yeah, definitely.

Yeah. Yeah. I have plenty more questions now, actually. Yeah, me too.

So do you want to start? Go go go for it, Jones.

Yeah, no, I mean I want to I want to first come a bit to the product perspective of things. It's been obviously quite a journey, and the product has evolved a lot over the last years, and now with new technologies opening up. Do you see this as kind of like a natural evolution of the product? Or were there moments where you felt like it feels a bit more like a pivot? And we all know, right, like the outside perspective and the inside perspective often is very different. But like, how do you go about this from a really like building a product up from the ground perspective? Is this an evolution, or do you feel like it's different product lines and pivots, even though if it's not a full company pivot, obviously?

A phrase I again want to stall but use a lot is like never mistake a clear line of sight for a short distance. And I, you know, we started the company with a very clear line of sight. Like when you're in the business of willing the future into existence, you know, kind of pulling the future into the present, you get to say what you think it looks like, right? But I think that's the most important thing from a product perspective is you have someone who has a very clear understanding of like if we do this correctly, it'll be like this. And that same someone needs to be able to storytell that vision. Whether you're inside a company like Adobe or whether you're a founder that needs to raise funding and convince people to quit good jobs to come and work for you in a grind, you know, you need to be able to storytell that vision in such a compelling way that people like I describe it like it, it has to feel like a quest. I I'll not try and remember the quote, but you know, Shackleton, you know, when he kind of did his big, you know, uh however the poster was for the job advertising uh advertisement, it was like, you know, we're gonna go to the Antarctic and we're gonna die. Who's coming? You know, and it'd be like, sign me up. So I think you know, you have to have a really clear vision of where you're going. And like there's a video that we shot for Assensi back in 2014. It was the first thing we spent money on, and it was the first thing we did. Then we did a pretty high production value kind of video of like this is what the future looks like when someone doing yoga or a kickbox or or a roar sticks this like you know, intelligent earpiece in their ear, and like, wait a minute, how can it see them? And it's coaching them. I watched that video back off and I'm like, we got it 90% right. And to your point about pivots, we've made massive changes along the way. You know, we were a wearable company before we were a computer vision company, and now we're adding coaching intelligence. And I'll pause for breath and let you decide where to go. But those were massive moments in the company, but I don't see them as pivots, I see them as like we had more information and so we had more clarity on how we got to that vision. Does that make sense?

A lot, a lot makes sense. And I I want to follow one question because you were saying that 90% of that video, it's almost real. It's like what you are doing. So, how do you manage the long-term vision on those days? Because in 2014 was a very different situation and build something was a complete different thing that we have today. Today with AI, you can build so many things, as you said before, with agents and everything. But how do you handle that day that you were building the long-term vision and the steps that you have to follow to achieve that vision? And nowadays you have more information than you make some tweaks. But how was the process for you?

Yeah, I when you talk about willing the future into the present, you know, it's got to be grounded in some pretty solid beliefs and understanding of what's possible today and what's going to be possible tomorrow. At my heart, I'm a technologist. You know, I in my time off I'm learning about technology and like all over the place, you know, robotics and space and yeah, I just it's fun for me, right? And so and I'm surrounded by founders who are the same. We have a kind of depth and breadth of knowledge about technology. And then if you back up to I joined Assensi from Microsoft, to its credit, Microsoft has incredible RD teams where you get to see the future. And so working with Microsoft, there was two things in all my work at Microsoft was all about an idea, a future, where we interacted with agents. This is 2011, 2012, right? I didn't know what LLM stood for. It was probably just research at the time. But we still allowed ourselves to imagine because I was working with teams like Microsoft Connect, so that's where all this a machine can see the world around it, and that's context. And I was working with teams like Cortana. Now Cortana, everybody thought was like trap surre, but Cortana was really a bigger vision around interactive voice models. And Xbox, Xbox Fitness, and Cortana, those three things together. And then also at the time, that was at the time that Under Armour were acquiring connected fitness companies, left, right, and center, and it was Nike Plus and it was Fitbit. And so I was in that whole world of we were seeing the very early heart rate monitors getting embedded in shirts and embedded in bras. And so put all of those things together, and that's where my clear vision came from. I I kind of saw all these things from afar, all moving towards this point of like, this is going to be possible. This thing I've always, this itch I've always scratched of like you can't teach karate from afar, and you can't teach any sport from afar. I think it's solvable. So uh I'll just kind of summarize it like that. You have to be able to kind of see long-term trends. But you actually you said steps, Marco. I again another kind of mentor, we consider a mentor, Jeffrey Moore, that I worked with at um Adobe and Microsoft. He always talks about the stairway to heaven. Like if you're creating a new category, you have to know what's the stairway to heaven. Like you've got to prove this, and then you get to do that, and if you do that, then you get to do that. So you have to have a clear uh stairway to heaven. It's very fortunate that AI and LLMs become true because without them, I didn't know how to do the last part of that video, which is like how can I have this like interactive conversation that just sounds and feels natural and is full duplex, which is just the nerdy way of saying we can interrupt each other and we can kind of talk very naturally. Now that's possible. Voice technology and LLM technology gives us a glimmer of hope that we can deliver on that last mile of the Essenti experience. So uh yeah.

That's perfect. That's great. And I love the the concept of obviously boiling it down into little chunks in that sense. And actually, it's something that I've been saying to the team a lot, and Marco will be like, here he goes again. It's like a product alone is not making it a company a business, right? And obviously staying in business for 12 years by now means every piece of that stairway to heaven actually needs to make sense in its own, right? Like it has to be able to stand on its own. Because if it's not, you're never gonna get to that next stage. So like you have this long-term product company, which I would call even more company vision rather than product vision, because the product is just the mean to get to the company vision. And now you're boiling it down into each and every step. But like, what is your framework? What is your thinking around how can I boil products really down to a position where it can make a business, at least for the time being, until new technologies enable the next maybe final stage right now is the LMMs. I I doubt it's the final stage, but like no, like no, we've got robots together. I think you're you're getting what I'm what I'm trying to get at. So, like, how do you think about boiling products down to a degree where it's actually commercializable and you can build a business around it that sustains the company for the time being until you can make the next step?

Yeah, and uh, you know, when I was with Adobe, I helped build Adobe Consulting. And so, you know, Adobe Consulting was incredible. Like we got to walk into customers like Nike, the NFL, NASA, NATO, all the ends. We got to like walk into these customers and deeply understand their problems and then kind of go back to base, if you like, and think of all of the technology we have available to us, ours and our partners, how can we solve that problem? And I was incredibly lucky to first work with and ultimately to lead the design organization, the UX design organization, the customer-facing UX design organization, not the team that was building Photoshop and uh building Adobe products. But what that allowed us to do, and something very novel that we introduced the industry to, and I'm very proud of this work, was a design-led consulting methodology that we would lead with design. This was at the time of companies like IDEO were kind of introducing the boardroom to the idea that you know designers should have a seat at the boardroom table. And design thinking is all about how do you have empathy for that person at the very end, the person that's sitting down and using the product, not the person that sold it to them, but you know, it's one step further. You'll hear me all the time in internal meetings talking about our customers, customer. Like we have to solve problems for our customers, customer, and we're doing that through our customer.

You must be suffering so much. So you being in that position as someone coming from design, having to trust that your partners and clients are doing a good job on bringing the experience to life.

Well, but the way that works, though, is we try not to present as a technology company. I mean, we couldn't be more deep tech. We're probably one of the deepest tech companies in our corner of the industry on the cutting edge of on-device LLMs and cutting edge of computer vision. But we try and not talk about that stuff so much. We try and talk about what's it like to be somebody that walks into a dojo for the first time, intimidated, and you might walk out again before the class starts, right? Or what's it like to train for a black belt? What's it like to want to do yoga, but you don't feel like you belong there yet? But that's the very reason why you should be going. And so we really try and design thinking is like, how do you live a day in the life of that person? And how do you uncover what are their unmet needs, what are the frictions or what are the things that you want to solve for them, and then bring that back to your product and like of all of the technology we have today, not what can we build in a year's time to meet that need? How do we meet that need today? And so I think our product roadmap has kind of naturally evolved with we always, you know, a phrase I again, anyone that knows me is rolling their eyes right now when I say art of the possible, but we always like to kind of show our customer this is the art of the possible. If you come on a journey with us, one day we'll be here together. Wouldn't that be amazing? Like come on the quest the same way our investors did, the same way our staff members did, come on the quest. But then once you come on the quest with us, let's find that quick win. What can we do in a quarter? And where could we be in six months and where could we be in a year? And with one customer, that informs your product roadmap. And if that customer matches your ideal customer profile, like if you get out of that startup mode that we've all been in, where like you'll do anything for anybody if they'll pay you for it, you're just trying to keep the lights on, you know, we'll code for food. But once you start winning your ideal customer profile, if you solve a problem, if you identify an unmet need for their customer, you can guarantee it's an unmet need for their competitor's customer, and that competitor becomes your customer as well. And so I really try and kind of it's very natural in our team. It's not forced, it's not like some prescriptive methodology anymore. We just bring that deep design thinking. And you've met many of our team. Like we understand coaching, like we're passionate about coaching, but we just so happen to be pretty nerdy about the tech that can deliver it. But every single day we just think, we use the phrase a lot in the team what's the best imaginable human-to-human coaching experience for this? How would we do it as coaches? Uh, and then how do we deliver that with technology? And that's the next thing on the road.

I love it. I love it. Because at the end of the day, it's the most important thing. How can we bring value not only for our clients but also for the end user? It's how can we bring all this information to them? And again, it it has to be aligned with the long-term vision that you have for your company and your product and everything. So it's like a Yeah, it's how we help and we can how can we build better coaches? Because I think coaches are the ones who are building your personality. Coaches are who build so many things for you, your mentors. It's basically a coach. Your father, your mother, your you know, your professor in karate in Taekwondo and Jiu Jitsu and everything. So it's like a you are trying to simulate the same thing in your own product at the end of the day, also, too.

So and I think really important as well on your roadmap is knowing what you're not going to do and why you're not going to do it. You know, it's uh you know Jeffrey Moore, who I mentioned, he would speak about core and context, knowing the difference between core and context. And core is what's the thing that you're known for. And people know that you do that better than anybody else, and the better you do it, the more value accrues to your company. And context is everything else. And if you take on that everything else, you might do it well, but you won't get any incremental value to your company because you did it. But if you do it badly, you'll be punished for it. And so, you know, knowing what's context and just and listen, I'm saying this on a podcast, it's like, oh, that's really good advice. That's right. I never follow it. My co-founders have to force me to follow it. I'm very bad at like, well, we could do this and we could do that. It's kind of the entrepreneur's curse, right? I was about to do it.

It's part of a dreamer. It's part of a dreamer. We are doing it.

I would just pretend I'm really good at doing that. But no, that's a that's a very healthy tension we have in our leadership team. Is there's people like me on this show that are here, like, oh, we should just do this. We can't it's easy, we only need to do that. And then there's you know, Ross and Bill who are like, Stop, why are we doing this? Who's gonna pay us for it? Like, do we have any evidence that the market wants it yet? What are you gonna drop? Like, that's what that's always a question to me. Like, if you want to do that, what are we gonna not do to do this instead? And I think you always have to be making these trade-offs of like what's not core.

You took away my question almost. It's like I was I was about to ask like stakeholder management, obviously, and coming up with the no list on the backlog, you know, like what we're not gonna be, what we're not gonna do, how do you go about this, trade-off decisions, etc.? I mean, you walk us through a little bit, like what's the process when you really, you know, you hit the wall and it's like we need to make a call here, like where where things go and you how do you make those decisions?

Is that I was thinking about this quite a lot the other day, and it you know uh and it's another martial arts framework for me, right? Because you know, the the metaphor for me is like we Are attacking the market. Like we're going after the market. Like we see it and we want to get there first and we want to get most of it. Like, don't underestimate the competitor in us. You know, we're we're going after it. And you know, again, this the martial arts framework for attacking, if you like, there's three there's gono sen, sen no sen, and sensen no sen. Gono sen is like you attack from a state of waiting, so you're reactive, right? You know, if somebody comes at you and you block and you counter, but yeah, it wasn't until they committed to the attack that you think about what are you gonna do about it, right? And that's how you start training, that's how you learn. It's block counter. You know, Marco, it's like you get on your back, somebody's gonna get on top of you, and now you're gonna learn how to kind of flip them over. And that to me is a roadmap discussion that's reactive. Customers came and asked for something, you're like, oh, we don't do that. And so now you're in a reactive mindset, right? It's like, well, how much time is it gonna take? And do we have the resource to do it? And how much are they paying for it? And does anybody else want it? And so I don't like attacking from a state of waiting. Like if you're doing that, you're always on your heels, quite literally. So let's go to the next one. Send no sen is kind of simultaneous attack. You know, you watch two kids at a karate competition, and they're just both punching each other constantly at the same time. Yeah, it's literally like that. And the positive way I would look at that is at least there's initiative, right? There's some initiative there. And I think about that phase as we're interlocking our roadmap with our customer. And we do a lot of this. We sit down with our customer, and it's like, now that you trust us, now that you know that our stuff works, now that we're partners together, let's open up a little bit here. We're all under NDA. Like, what does the next 12 months look for you? What things are you thinking about? And let us show you our roadmap and how we're thinking about our roadmap and react to it. You know, and we're very customer driven on our roadmap. It's like if you tell us that that thing we're going to do in 12 months, you would use it now. We'll pull that forward and we'll push something out. And when you start doing that with multiple customers at the same time, your roadmap starts to interlock with the direction the industry is going in, not just the direction the company is going on. So that's a really that to me is the that should be the steady state that we're constantly roadmap interlocked with our customers. But the last sen no sen, that's like to show initiative before initiative. Have you guys seen The Last Samurai? You know the movie with Tom Cruise? Yes. There's a move, there's a moment where he kind of gets jumped in the town at night and they play it in slow motion first, and you you see him observing all of those things kind of almost like before they happen, and then you see the combat, and it's like, and then he knows where everybody's gonna be before they know they're gonna be. That's sen sen no sen. And what that really is is acting on intuition. You're picking up on things that because you're seasoned, because you're so you've got some element of mastery at what you're doing, you start picking up on cues, and so you can see things before people maybe know they're gonna do. I could talk about this for days in the context of you know sport karate and martial arts. But I think every so often you have these moments in the industry as well. And coaching intelligence is one of those moments for us. It was like there's no guarantee that the industry is gonna move towards on-device LLMs instead of just calling open AI and calling anthropic. But we held that belief two years ago, and I'm giving myself too much credit here. My head of AI held that belief two or three years ago. He's like, it's all gonna be on device, you know, it's not gonna be in the cloud. And over the course of those last few years, we've collectively gathered more and more conviction that like this is right. There's no way people are going to be able to like offload their business to Anthropic in the cloud. Make me a training program, upload a video, tell me what I'm doing wrong. Like that's just not for any token costs and privacy and security and latency. And so we had this increasing conviction that on-device LLMs, on-device voice cloning, the model will be less important than the harness around it and the memory that has and the context that it has. And then there comes a moment where you initiate the attack. You don't wait to be attacked, you initiate the attack on intuition. And those are dangerous moments, you know, because if you get that wrong, if you weren't going to be attacked or there was nothing to attack, you get that wrong. But when you get it right, I think it's company changing. And I you know, I just feel like we're just on the other side of one of those moments that essentially of like the bet on coaching intelligence will be way bigger than the bet we made on getting movement recognition right. Because every single health and fitness brand is going to have an AI coach, and they can't afford the tokens to run that on open AI frontier models or anthropic frontier models. It'll have to be their model running on a device, you know, running on the consumer's hardware. So yeah, that's how I think about you know roadmap.

I love it because I and I love how you explain it because I do believe it's a combination of everything. It's like at the very beginning, also you are trying to we said before, the vision. You are trying to, with all the things that you see around, you can build the vision and then you start fighting and showing, and then you will defend a little bit because at the end the feedback that you are going to start receiving is hard at the very beginning.

It's bugs.

Yeah. Exactly. You have how many bugs do you have in the features? So it's like you are like this, and then again you are attacking, and then you are again showing like, oh, I can see these spots here, and then it it's a whole process.

You just gotta be in the mix, right? I mean, you can't stand back and watch. You just gotta be in the mix and you gotta trust that you and the people around you. I love I love that saying that you know you have all the resources available to you or the ability to harness them to do anything. And so you just have to kind of have that trust that you know we've probably got the resources we need or we can go find them to do this.

I agree. Be water, my friend.

Which is a great note to to kind of come to a closing. And we always have to ask the one final question because the podcast is called This Feature Will Save Us, and I have a bit of a feeling already, like But yeah, what's the feature, Steven, that's gonna save us?

You know, I uh you know my latest provocation, if you like, is longevity is causing a labor crisis. So as we live longer, Parkinson's is gonna double by 2050. There's gonna be 80% more people having strokes in the 2030s. And it's nine out of ten of those people having strokes will be because they lived longer. It's not because the environment's changing or you know, it's quite simply the longer we live, the more likely the odds of something that would happen to us happen to us. You know, one in two people after the age of 50, I think, have a you know bone density related fracture. Well, the more people live after 50, the more people are having bone density related fractures. And then if you look in the fitness industry, you know, the fitness industry recognizes you know GLP1s and you know the drugs that are coming behind them, amazing for getting people back to wanting to be physically active with the recognition that you shouldn't be prescribing these drugs unless you're strength and conditioning training, you're building you know lean muscle. And so that's more people that want to enter the fitness industry. But you know, when you look at the labor market today, if you look at some of the stats, you know, most fitness trainers burn out after two years. I think 80% burn out after two years. And I think I saw a stat recently that for every person in a fitness instructor job, there's three open positions waiting to be filled. There's not enough strength trainers, but it's also it's not affordable and accessible to everybody. Like not everybody, like buying a gym is a high consideration, buying a gym membership is a high consideration purchase for a large percentage of the population. Tacking on personal training on top of that is not even a consideration. And physical therapy, it's worse. You know, 72% of PTs in the US have more demand than they have time. And if you live in rural America, it's twice as bad. You know, access to PT in rural America is twice as bad. And burnout in PT is real. It's like high 70% of PTs report clinical burnout because it's just such a high demand job. So we're not hiring our way out of this problem. So what's the feature that's going to save us? It's an agentic workforce. And we're already seeing that. You look at companies like Eleven Labs, you know, they're manning the front desks, they're making and taking calls. We've got agents that can work alongside us where the phone is ringing off the hook and nobody's answering it. Now an agent can step in alongside the human workforce and do that. We're seeing it in legal work, we're seeing it in accounting. And I believe there's no reason we won't see it, and we have to see it in health and fitness. We have to support the humans that are tapped out trying to provide care and coaching. We have to support them with an agentic workforce. And so that's the feature that we hope is going to save this small part of the problem is if we can give a machine the perception to understand human movement and we can give it the ability to converse in a knowledgeable way and guide you through your workout in the gym or guide Marco through his ACL recovery, or help my dad when he had a stroke during COVID and he had to do his occupational therapy at home, you know, AI can save us there. So uh that's my hope and uh it's an exciting journey to be on.