5 Keys to Amazon Alexa's Conversational AI with Emerson Sklar

Emerson Sklar, Chief Evangelist of Amazon Alexa, joins host Dr. Joan Palmiter Bajorek on Your AI Roadmap to discuss the 5 Keys to Amazon Alexa's Conversational AI. Emerson explores how large language models (LLMs) are enhancing Alexa’s ability to be more conversational, useful, personalized, trustworthy, and engaging. He also delves into the evolution of Alexa, from its consumer-focused origins to broader enterprise applications. 

Podcast:

🗳️ Nominate Podcast Guests! https://yourairoadmap.com/podcastguest 

Quotes

🌟 "The North Star of Alexa has always been to build the best personal AI assistant, something that's as natural as talking to another human."

🧠 "One of the challenges of large language models is they're great at knowing things, not great at actually doing things, and actually acting on the world on your behalf. But Alexa's fantastic at that.”

🔒 "But people rely on Alexa millions of times each day, and we need to make sure that we aren't introducing any kind of behavior that breaks that trust.”

🎤 "I know some brilliant technical folks that are good in front of a room full of colleagues, but the idea of getting up in front of a stage and talking to hundreds or thousands of people is be paralyzing"

🛠️ "Amazon has this notion of leadership principles that are sort of guiding tenants… One of my favorites is customer obsession.”

Emerson Bio:

Emerson Sklar is the Chief Evangelist of Amazon Alexa, where he helps enterprises and developers around the globe realize the benefits that conversational AI can bring to their customers. He is a recognized thought leader, public speaker, and recipient of the 2022 AI Excellence Award. He has spent his whole career helping companies improve their apps, devices, and experiences with a user-centric focus on quality. Before Amazon, Emerson worked at Applause, Bespoken, Borland, and the Army Research Lab. Outside of work, Emerson can be found volunteering for Scouting America, relaxing with his family and flat-coated retrievers, and making unnecessarily complex meals.

Connect with Emerson: https://www.linkedin.com/in/emersonsklar

Chapters

• 0:00 Introduction and Overview
• 1:51 Evolution of Amazon Alexa
• 3:44 Success Metrics for Alexa
• 4:13 Why Amazon is Best Positioned
• 9:14 Surprises and Challenges
• 17:50 The Role of Alexa in User Adoption
• 17:53 Emerson's Career Path
• 21:50 The Role of an Evangelist
• 26:03 Building Presentation Skills
• 29:06 The Importance of Mentors
• 31:00 Getting Started in the Field
• 33:59 New Chapter

Transcript

0:00

Hi, my name is Joan Palmiter Bajorek. I'm on a mission to decrease fluffy hype and talk about the people actually building in AI. Anyone can build in AI, including you. Whether you're terrified or excited, there's been no better time than today to dive in. Now is the time to be curious and future -proof your career, and ultimately, your income. This podcast isn't about white dudes patting themselves on the back. This is about you and me. and all the paths into cool projects around the world. So what's next on Your AI Roadmap? Let's figure it out together. You ready? This is Your AI Roadmap, the podcast. Hey folks, this is Joan jumping in to share a little bit about the episode. This episode is with my friend Emerson who works at Amazon and he and I have known each other since, my gosh, 2018 where we met at Voice Summit and shortly after he hired me for some contract work and we have been colleagues in this field of voice and conversational AI since then. regardless of what companies we're working at, I think of him as a colleague. And I really love in this episode how he walks through best practices, thinking about the Amazon Alexa product, but as well as how we think about agents and conversational AI when we interact with different systems. I think you're gonna really like this episode. Emerson can't share everything that's being created and deployed, but hopefully it opens your mind. I think there's an example with a garage door that is so practical. I hope you listen in and really enjoy. Hello, hello! Howdy, thanks for having me. Absolutely. Could you introduce yourself, please? Yes, my name is Emerson Sklar and I am the Chief Evangelist of Amazon Alexa. Awesome. All right, Emerson, what are you working on right now? What aren't we working on over at Alexa? I think the big thing, you know, last September, we gave a preview of the future of Alexa that is powered by some, a new LLM and a suite of new and enhanced kind of conversational AI capabilities. And the North Star of Alexa has always been to build the best personal AI assistant, something that's as natural as talking to another human. And the advent of LLMs, I think, really allows us to realize that at North Star. That's awesome. Well, and for those, I mean, everyone on here is gonna have heard about Amazon Alexa. But when you think about, and you're talking about a suite of tools, I'd love to hear just about what you think the evolution or how you all talk about the evolution of like Amazon Alexa's first deployments back in what, 2016, 2017, to what it looks like today in 2024. I still think about it as a customer device. But the way you're talking about it, maybe it's more business context. how do you think about that evolution? Yeah, I think a little bit of both certainly. I mean, when the original vision behind Alexa was to have the Star Trek computer, this omnipresent, all-knowing, all-capable machine that you could just talk to on a wide array of topics. And when Alexa was first announced in 2014, I think the first device, the first Echo came out in 2015, when those came out, The focus certainly was on the consumer aspect, on the consumer interaction with these smart speakers. And Alexa really effectively defined both the common notion of a consumer interaction with an AI assistant, as well as that notion of the smart speaker themselves. What we saw both initially in a sort of immature kind of fashion. And we've seen an evolution over the last almost decade, growing maturity and growing focus on is some of those more enterprise centric kind of business cases that provide still significant consumer value or consumer engagement or entertainment or anything like that, but then have some real meaning and impact on a brand or an enterprise itself. Even to this day, we released one of my personal favorite of the Echo devices, the Echo Show 15, which is the large rectangular one that I have a bunch around the house, but we're constantly identifying new form factors or visions of existing form factors, and they are extraordinarily popular and ubiquitous in the dozens of countries that Alexa supports today. Absolutely. When you think about kind of metrics of success that you all are talking about today, What does success look like in these next stages? Yeah. So we haven't revealed much publicly yet. We did have the initial announcement back in September of last year, and we had some additional announcements at CES this January. But generally, when I think about what we're doing, the internal guidelines of the measures for success are around five key characteristics of Alexa. both Alexa in the past and then Alexa certainly moving forward. So one is that we want it to be as conversational as possible. Alexa already has incredibly powerful ASR, automated speech recognition, and NLU natural language understanding. But it doesn't perfectly understand everything that every user says. And research has shown that you shouldn't try and train the user on how to speak and interact with these kind of devices. You should. as best as possible support the natural way that they're going to do it. But that hasn't been trivial for developers to do effectively in the past. And so the advent of LLMs, I think, radically improves that conversational understanding. It makes it as well, kind of much more able to respond back to the user in a human-like way. You know, think about like this morning, I woke up and thought that it was pretty dark and cold outside and asked if it's going to rain later today. And so when I ask that, I probably just want a yes or no. I don't care about the rest of the weather. I just really want an answer to that question. Versus if I said, oh, what should I wear today? Then perhaps I'd get a more extended type of answer based on the weather. There's also a ton that goes into a conversation beyond just the words. There's body language, eye contact, there's context that you know about the person that you're speaking with. And so for it to be. really conversational, sort of that next generation level of conversational, it needs to be able to understand and interpret kind of all of those signals. So conversational nature is the first one. The second is that it's got to be super useful. And one of the challenges of large language models is they're great at knowing things, not great at actually doing things, and actually acting on the world on your behalf. But Alexa's fantastic at that. Last I heard, customers have connected more than 400 million smart home devices to Alexa. And so imagine every day I get home, I might do the same thing. I might open my garage door, drive in, park, turn off the lights, disarm security system. With Alexa today, I could try and set up a routine that would do that automatically for me. But it's going to be tough, and it's going to require me to try and put in that complex logic in an app on my phone, which I might be able to do and you might be able to do. But maybe our parents would struggle with it. It'd be really useful if I could just do that with natural language. So I could say, hey, Alexa, every afternoon, once I park the car, close the garage door, turn off the light, and turn off the security system. Um, so making it then not only more capable and able to do things, but easier to get that kind of result, uh, and action that you want is, is critical. That's the second thing, useful. I think the third is it should be hyper-personalized. I've said for a long time, my experience with an AI assistant should be different than your experience, should be different than the listeners or watchers, should be different than my two-year-old's experience. And Alexa has some of that personalization today, but it's not really consistent across experiences, and it's not particularly comprehensive. Just the other day, I said to my wife that I wanted to order dinner, and she knew exactly what I met. She knew that I wanted Chinese food. She knew the restaurant that I wanted to order. She knew exactly what kind of food I wanted, what customizations I wanted. And so a more personalized Alexa should know all of that as well without me having to go through some kind of back and forth to do that every time. And that requires a drastically expanded context as well. So if I said, you know, give me... this order and I said it was too spicy last time. It should know that like too spicy means the five star General Sous Chicken that I got, and it doesn't mean the one star wonton soup or something. Again, sort of just as a real person. So that's personalization. I think personality is critical. Something that differentiates Alexa from other voice assistants is its personality. It's friendly, honest, factual. But it perhaps doesn't emote the same way that a real human would. I'm not a sports fan, but imagine that I were a sports fan if my favorite team like just want a really big game. I want Alexa to be just as excited and celebratory as I would be. If they lost, I want it to be a somber but optimistic, like, oh, look on the bright side. If I ask an opinionated question, like who should have won the best picture this year, I want it to give me a friendly and strong point of view in return. And so the approach that we're taking to LLMs enables us to maintain kind of those characteristics that people know and love. while still enhancing and improving the experience to feel much more natural and engaging. And then the last success metric is being trustworthy. And perhaps the most critical, I mean, everyone has seen the sometimes comical, sometimes really harmful hallucinations that LLM-based systems often exhibit. But people rely on Alexa millions of times each day, and we need to make sure that we aren't introducing any kind of behavior that breaks that trust. We need to make sure that we maintain factual accuracy, that we protect customers' privacy, we protect their data security, and maintain that trusted AI assistant relationship. So those are all big challenges. And we have massive teams working on each of those pieces. But I do think that Alexa is perhaps better positioned than really any other company on the market to try and solve those challenges. wow. Ambitious, concrete. You all, I'm always impressed. One, why do you think you all are best positioned? So Amazon has this notion of leadership principles that are sort of guiding tenants, both for how we operate as a business internally, how we interact with our coworkers, as well as how we create new products and deploy those products out to the market. And officially, there's no more important or less important leadership principle. But one of my favorites, is this notion of customer obsession. And so rather than starting and saying, OK, our competitor is doing this, so let's fight against that or let's copy it or something, instead of doing that. And instead of saying, oh, hey, there's a gap in the market. Let's create a product to fill that niche or something. They start from the needs and wants and expectations of the customer and then work backwards from there to try and find some sort of solution. And we've been doing that, I guess, I mean, Amazon's been doing that since long before Alexa was around. But it really is fundamental to Alexa business and deep in the DNA of Alexa. And I think that history of understanding what those customer needs are and how we can build the best voice assistant. the technology and the industry to get to where our expectations and our customers' expectations are. And the fact that we have just breathtakingly brilliant scientists and designers and engineers and business people who are super passionate about what we're doing, I think that puts us in an incredible position. Certainly makes it an exciting and engaging place to work. Oh, for sure. Well, and as I think about kind of where devices are today, and as you mentioned, like, some people wish it could do this. Like, why doesn't it already prep my coffee in the morning and blah, like this personalization, these beautiful user flows. But also, as you and I know, connecting different APIs and different tech, and there's one little breakdown, and then it's not going to function. But that's today. And what could be? And especially when I hear, did you say 400 million? Over 400 million, I mean, just the size of these data sets, Emerson, whole like, whoa, my mind, whoa. billions of interactions every week. I mean, it's incredible. Yeah, that's a lot of data. Well, we're speaking with a Googler on this podcast. We were talking about just the size of their datasets. And she said something akin to like the whole internet that we're allowed to touch. I started my career in government contracting, working for the Army Research Lab and then the Army Intelligence Securities Command and then did government sales as a solution engineer for a number of years. And Amazon takes security and privacy more seriously than some top secret facilities that I've been in. And I don't know if that said something positive about Amazon or something negative about the state of our government security, but it is both top of mind and kind of a non-negotiable critical part of how we signed Alex initially and how we are continuing to design in the future. I mean, any of our devices, like I'm speaking to you from an iMac, like all these devices have more and more listening capabilities and how our data is protected, what different companies, government entities are doing. It's a very fragmented picture. When you think about, I love asking this question because I've heard... The answer is I was not anticipating. When you think about surprises and challenges that you have learned across the last few years, are there any things that you have been like, whoa, this came out of left field, was not expecting, or are any surprises as a practitioner? Yeah. So I think one of the biggest surprises is how much LLMs have reinvigorated consumer and end user and layperson interest in conversational AI and voice assistance. the natural reaction 14 months ago or so when ChatGPT came out, it's like, Uh-oh, that's a big deal. That is quite a new entrant into the AI assistant space and capability. And what's amazing, I mean, it took ChatGBT less than two months to hit a million active users, which is the fastest growing user base for an app ever. You might look at that and say, okay, well, you know, that's a million people that are using that. That's got to be a million less people that are using something else that might have existed before. It's very much not the case. Again, people already use Alexa billions of times each week, and that utilization continues to increase by leaps and bounds. Again, I think partly driven, obviously, by the continued innovation that we have, by the new devices, by growing comfort with it. But but also driven by some of the additional interest and excitement that other players in the space have introduced. So that was an exciting surprise, I guess. On the challenge part, one of the challenges for me personally is staying patient. So part of my job. as the chief evangelist, is to evangelize, is to go out and publicly share and build excitement for what we're doing and show off cool new sci-fi stuff that we have. And there is so much cool stuff that I am lucky enough to get to support and to get to see and interact with. But it's not my decision about when we get to showcase it. And so it's... It's personally tough to keep a lid on all that really amazing innovation. But on the flip side, the other half of my job is being the internal evangelist. So meeting with our partners, our users, industry thought leaders like you and some of our friends and serving as their voice and advocate internally to help ensure that we remain customer obsessed, that we stay aligned with the needs and expectations of our broad partner and user community. And that part is, I think, if anything, even more critical at this stage and has absolutely kept me busy for the last year plus. That's awesome. Yeah, well, I think that patience, you are not the only one I've heard this from. My friend working, or yeah, Accenture met, and like I worked on this for three years and I can't tell anybody, like anybody what's going on in the lab. Cool, well, that's some really exciting stuff. when we project out for our field. And you and I have known each other for a fair amount of time now. We met at first in 2018 and like thinking about where we thought the field was headed then. you know, and doing a check-in here in 2024, when you think about the big picture of this side of the field, of LLMs, of smart devices, connected devices, et cetera, be on Amazon, this is your answer, not Amazon's. When you think about... Right, right. Not trying to get anybody in trouble. When you think about those bigger pictures, what do you see as the next steps? Yeah, good question. So for AI assistants in general, certainly see further embracement of both large language models and large action models to provide much better knowledge, utility, conversational interaction. And I think that staying on top or staying at the forefront of the AI system the evolution of that technology, figuring out how to properly implement and manipulate it is critical. And so it's something that everybody is going to need to get much more familiar with, again, as quickly as possible. And these new tools really do remove a lot of the tedium of traditional interaction model-based conversational app development. But it adds a significant challenge to figure out how you can craft and shape the experience to accurately reflect. whatever your company's brand and personality and reputation would be. So people say a lot of the AIs take jobs away. In some cases, perhaps. And in this case, it may shift what that focus is, but still a tremendous amount of work to be done to make sure that you can create some kind of AI-based solution that reflects your company's ideals. I think we, it feels, where we are today feels to me very much like in 2015 and 2016 when Alexa was first out there and suddenly every company said, all right, we need a voice app. And then they go and they lock a couple of people in a room for six months and have them do something and they don't really do their due diligence to figure out what's actually meaningful for their business. And then they release it and it gets terrible reviews. and then they swing the other way and become far too conservative with future investments or future developments with the technology. It feels like we're in a similar spot now that everybody's experimenting, everybody is rapidly prototyping and just testing things out and tinkering. And I think that is great. But I think we will see in the near future a shift to really getting down to much more useful, perhaps more constrained, simple, but again, still useful kind of use cases. I think there's an interesting unsolved challenge right now of resource optimization, the computational power, and then the dollar cost of the system. anything LLM powered, supporters of magnitude more expensive than traditional voice and chat experiences. And so some of that can be ameliorated with improvements in hardware, right? Some of it can be finding novel ways to use legacy technology where it makes sense, okay, Emerson, let's switch to your career stuff. This podcast is called Your AI Roadmap, because so many people either are in tech and wanna pivot, they're like, heck yeah, AI, let's go. Some people aren't, and they're like, ooh, everyone's at a different pivot point. If someone hears this, your story, and is like, ooh, I wanna go on Emerson's path, how did you get to where you are now? Yeah, it's not a direct path. So my first real job out of college was an army contractor. And we were building these mobile supercomputers that we could drop ship anywhere in the world to provide kind of local high performance intelligence analysis capabilities. And that was my first real professional foray into AI. It's such a tremendous amount of data at our fingertips. uh, these systems were able to provide insight that a human could never possibly do. I mean, it, uh, it's the, the finding the needle in the stack of needles, uh, kind of analogy. And, uh, I, on, on that project, we, uh, contracted with somebody called Borland we leveraged their software development life cycle optimization tools. And when we signed that contract, I stepped up and led a project to implement it. Um, and I met my first, uh, really extraordinary professional mentor, um, who ran the Borland federal team and he made an offer to me to come be his solution engineer sidekick to then help. advise and implement the software for federal agencies all over the country. And so that was a really amazing opportunity to get exposure to tons of innovative use of technology. Not all AI powered, but really across the spectrum and was a great way to really deeply dive into all parts of the software development lifecycle from conception to deployment and maintenance. spent four years there and then joined a number of my former Borland colleagues over at Applause, who is a leader in crowdsource testing and user research. That's a fascinating company because they kind of aligns with this notion of customer obsession. I was fortunate to meet, to gain other really incredible professional mentors there, one of whom managed the Amazon relationship for Applause. They did some business with Amazon at that point. But when... Alexa first came out, he said, hey, I want you to learn this stuff. Trust me, you're going to enjoy it and it's going to take your career off in a different direction. That was my first real introduction to the modern world of conversational AI. He was right. I remember working hand-in-hand with a ton of the early Alexa team. There's a big developer advocacy group as well. I remember seeing what an incredible drive they had, the sense of innovation that Amazon had cultivated, their enthusiasm for this new technology, and the really industry-shaping impact that they had on everyone. saw that and said, I want to do exactly that. That's what I want to do. I had no idea how to get there or what to do to follow that path. But again, I was fortunate that I had some really extraordinary mentors who helped guide me on what I needed to do, what I needed to learn, the skills and experience I needed to have to be able to get into a position that I would be useful for Amazon. in such a role to actually be considered for it. I was at Applause for almost five years, and we had a ton of interesting partnerships with Amazon, helping to collect data, train new voice models, testing pre-production devices and software. And that part of the business grew enough that I founded the Applause voice and AI practice, which then grew exponentially both with Amazon and outside of Amazon. So again, sort of. tons of really tangible hands-on experience with this kind of technology, but at a wide range of companies. So I was there for five years. I took a two and a half year break in the middle to help run a small startup called Bespoken, who are the leaders in conversational app automation, automated testing and training. And then Amazon opened up a new position or initially for the chief evangelist of Alexa Enterprise, which is one component of the overall Alexa business. And I guess the rest is my Amazon history. That's awesome. Well, and as you and I know a lot of evangelists at different companies, but if people haven't heard this term where they're like evangelist, is that, is that, it sounds like a tele-evangelist or maybe like an influencer or is that like a cheerleader for the product? Like when you think about the term evangelist in your role, how do you frame that? Yeah, I think it's twofold. So one is serving as the external face of the business to our partners and developers. So providing thought leadership, providing best practice guidance, real hands-on expertise and experience to help ensure that companies who are already using our technology are already adopting it, that they are as successful as possible, as well as then new partners that may not be using it yet, that they understand why they would want to and how they could do it and what the potential benefit would be. We're so close to it, right? It's so easy to just assume that everybody's experience with the technology is the same as ours and that everybody has the same perspective and the best way short of actually, I learned the value of that very early on. When I was at Applause, we were working with a big American fashion tailor. They were building Alexa skill that You could check your rewards points balance. You could see when the nearest store to you is going to be open. You could check the status of your order, like kind of basic stuff. And they had spent a bunch of time implementing this affirmation capability, where you could say, like, hey, Alexa, ask this retailer to give me a daily affirmation. And it would be like, you look great today. Or like, you're going to do awesome. Have an awesome afternoon. And I was like, that is the stupidest thing that I've ever heard. Why would you spend money and developer effort doing that? Nobody's ever going to use that. But what an odd thing to waste your time on. And, add applause, because again, we used real people. We had recruited a whole bunch of people that matched their target demographic, which was predominantly middle and upper class women, like ages 40 and above or something. It's a very different demographic than I am. And did a bunch of user research and they loved it. They loved the ability to have it give those positive affirmations every day. And it was really eye-opening. It was a real testament to how important it was that company really knew their users really knew their target demographic and their customer base, as well as then how unique people out in the wild, how unique their experience is and their expectations and interaction with conversational AI. That's again, something that I wouldn't have known had I not gone and helped lead that kind of study. Oh, totally data driven, right? Like asking, this reminds me so much of like, uh, I don't know if you remember early dash bot of like analytics tools. Um, and there was a presentation about like emojis, keep breaking the chat bot. I'm like, people responded with a smile. You guys sound like, it's going to person like, uh, but the number of times I've built out for, um, for different enterprise bots, like chit chat features, we'll asking like, what's the age or like, how are you doing today? And the bot. totally breaks. But if you make one of those 10 to 15% or just people just want to have little daily convos, it's shocking even for like a government bot. I was like, really? Okay, let's build it. Like, I know what's the problem of sending smiley faces or have a good day, you know? Yeah, and again, that's something that for the traditional way that people built voice and chat bots, you had to manually go in and try and cover as many of those edge cases as possible. You had to spend a nontrivial amount of effort to try and make sure that it wasn't just going to fall flat when somebody did one of those off-the-wall use cases. And it is a nice thing about LLMs is they really understand how people talk and how you write language. And again, makes it, still makes it a challenge to figure out, okay, with that great understanding of natural language, how do I respond to it? What do I do with it? But certainly provides the opportunity to provide a much better customer experience in the end. Mm-hmm, absolutely. Well, and as I think about your skill set, you talk about this solution engineering, finding mentors. I'd love to talk about that. But a lot of your role, I've seen across the years, very public facing, big presentations, talking, public speaking. And I actually, there is a, I don't know that it matters now, but Google Cloud offered me an evangelist opportunity back in the day. And I was like, I'm qualified. Um, my female friends who have technical evangelist roles, the amount of people questioning their Python or Java skill, or just constant public facing, no matter how qualified you are, have a PhD and then done this a bunch of times. Uh, you just, like, it's a, it's very big and like 30,000 people potentially attending some of your talks. Um, maybe that's not exactly what you do, but in those type of, if people heard this and they're like, oh, I like what Emerson's talking about, public facing, public speaking. Like, how did you build that skillset? How do you navigate that today? Yeah. I've always liked public speaking. I like the showmanship aspect of it. It, to this day, still makes me nervous, which is one of the reasons why I really enjoy it. It's not something that came naturally to me. I mean, again, I have a degree in computer science. When I started my career, I was doing full-time software development. And that was much more natural. It was when I went over really to Borland initially, and then suddenly had this new role where I'm... partnered with a mentor and friend of mine and having to get up in front of partners and customers. Smaller settings, but still intimidating settings, where like you said, I mean, at that point I was 22, 23. And yeah, they question, is this person credible? Does he have the knowledge and expertise to be able to? properly recommend and give advice to our massive business. And so it was fortunate. I mean, Borland had been around for a long time. At one point, it was the second largest software company besides Microsoft. So they had a pretty well-oiled pipeline of how to ramp someone up, as well as a bunch of incredible experts that had been doing it for 20-plus years. And it was a conscious effort to try and figure out how to do that. And it's not... Perhaps it's easier now, but it's no less exciting and no less anxiety-inducing in a positive way. That's the thing, right? It's not for everybody. I know some... brilliant, brilliant technical folks that are good in front of a room full of colleagues, but the idea of getting up in front of a stage and talking to hundreds or thousands of people is – would be paralyzing. It's not for everyone, but I do like that positive sense of tension that it gives me. It keeps me on my toes, I guess. I look at you, you're like nervous and a good thing. I feel similarly, and I've definitely seen in my career, technical skills get me in the door, but some of these like soft skills, presentation skills, leadership sees you, and that's how like technical skills put me in the door, soft skills are how I rise, is how I've experienced it. I think that being translatable, like your technical work. actually being consumed by who needs to hear it has been crucial. When you think about, so you mentioned mentors and especially as people think about their careers, etc. I mean, sometimes I think mentors is overused, like some person's gonna pick me out the sky and like save me or something like this and is a person an advocate, blah, blah. Like how do you think about, I mean, you've mentioned these people who are pivotal to your career. How do you foster those relationships? Did they pick you? Like, how would you talk about people who are like, Oh my gosh, a mentor! Oh! Yeah, I was fortunate that in I think in many cases, they had picked me and given me great advice. And I think it's, obviously, I had a hand in it from the position that I was in, from being open and receptive to getting feedback. But I consider myself extremely fortunate for the good friends and mentors and allies that I've had really, really throughout my career. People that, many of them are not at all technical, most of them know not much about AI, certainly not conversational AI, but that very diverse perspective from people who are skilled professionals, experts in something that they do in... technical sales to federal agencies in journalism and whatever it is, that diverse perspective really put me in good stead and was very valuable in helping me get to where I wanted to be, I guess. But that's very much not the case for all of the mentors that I've had and very much not the case, I think, for most people. It does, in many instances, take some really concerted, thoughtful, direct effort. And that can be uncomfortable sometimes. I can think of situations where, again, especially when I was fairly junior in my career, where I put myself out to someone who was much more senior and much more well-known or well-regarded, whatever that... It can be uncomfortable, but it's not going to happen without that. It's not going to happen without people knowing that you are receptive to or actively seeking or wanting feedback and support. I feel fortunate to be able to do that as well with people that I mentor. Right? It's... It is... it made such an impact on my professional and personal trajectory. And I think it's equally important than to be able to try and make that same kind of impact on other folks as well. Absolutely. Good. Before we give it back. And speaking of that, other than, you know, reaching out to folks, if people are trying to join our field, are there programs, you know, you have degrees, like internships, like how would you recommend people start taking those steps towards these careers? Yeah, so certainly for the technical evangelism and for any machine learning and development roles, having a technical background was very helpful. I have a degree in computer science from West Virginia. And most of my career, I've not been a full-time developer. Was at the beginning, and I've done some shade of part-time development essentially the rest of my career. But those... Those CS fundamentals, the technical and problem-solving capabilities have been super useful throughout. I had a good friend and colleague on my first real job that I initially was doing QA work when I first started. And I was really not enthusiastic about that. And at that point in my career, I was really not enthusiastic about it. And I had this friend and colleague on my team who saw that I was not so enthusiastic about it and said, look, you have to do that job. You have to do that role because that's what they're paying you for. That's what their expectation is, is that you're going to do it. But you should figure out how to do it effectively and thoroughly and efficiently. And then with your extra time, go do something else, something else constructive. And they're not going to be upset if you're doing something that's still furthering the project's mission or whatever that might be slightly outside of your role scope. And that led to me getting into test automation, which I love automation. That led to me taking on that project, leading that Borland Initiative. I mean, that really was the first building block for the upward. trajectory or the different trajectory of the career. So for people that are not in this field already or may not even be in a particularly technical field, I think there's so many tools, there's so many resources out there you can get started even if it's not your official current job description. There are lots of AWS published. I think like 50 different classes on generative AI and ones for business decision makers, ones for hyper-technical data scientists, for application developers, for everybody in between. So there's tons of great free learning resources and certifications that you can take. And then Yeah, and then just getting started and trying the real hands-on experience, I think, is critical and one of the best ways to learn and evolve and start becoming comfortable. Yeah, I couldn't agree more. Well, thank you so much for your time. If people wanna follow up with you, where can they find you? Yeah, you can find me on Twitter, on LinkedIn, under my name, Emerson Sklar. Certainly feel free to shoot me an email, esklar@amazon.com. Yeah, yeah, would love to chat with anybody who's already in the field or anybody who's considering getting into the field. Happy to continue the conversation. Awesome. Well, thank you so much for this discussion. I've learned a lot. You're welcome. Absolutely. Have a good rest of your day. Yeah, yeah, you too. Oh gosh, was that fun. Did you enjoy that episode as much as I did? Well, now be sure to check out our show notes for this episode that has tons of links and resources and our guest bio, etc. Go check it out. If you're ready to dive in to personalize your AI journey, download the free Your AI Roadmap workbook at yourairoadmap .com / workbook. Well, maybe you work at a company and you're like, hey, we want to grow in data and AI and I'd love to work with you. Please schedule an intro and sync with me at Clarity AI at hireclarity .ai. We'd love to talk to you about it. My team builds custom AI solutions, digital twins, optimizations, data, fun stuff for small and medium sized businesses. Our price points start at five, six, seven, eight figures, depends on your needs, depending on your time scales, et cetera. If you liked the podcast, please support us. Can you please rate, review, subscribe, send it to your friend, DM your boss, follow wherever you get your podcasts. I certainly learned something new and I hope you did too. Next episode drops soon. Can't wait to hear another amazing expert building in AI. Talk to you soon!