Michelle Hamilton from AnswerRocket Leads AI Adoption That Actually Works

Michelle Hamilton has spent almost 40 years as a glass sculptor. She's also the Director of AI Adoption at Answer Rocket, running enterprise AI rollouts for Fortune 50 clients. On this 80th episode of the Bald Ambition Podcast, Mookie sits down with her to find out how those two careers connect and reinforce each other. 

Both jobs, Michelle argues, come down to the same skill: visualizing a client's vague need, walking them through it verbally and mentally, and translating "what's keeping you up at night" into something concrete. She's currently engineering an 8-foot exterior glass sculpture — over a thousand melted glass pieces, lit from within, built to survive Midwest snow and rain — and she did it by uploading her hand sketches to an AI and working through the full engineering scope with it as a thinking partner. That's the same kind of conversation she has with enterprise clients trying to figure out where their AI investment went.

And that's the real subject of this episode: AI capability, sure, but also AI adoption: the much harder problem of getting human teams to actually use the tools their company already paid for. Michelle's central finding, after two and a half years doing this full-time, is that the technology is rarely the bottleneck. Fear is. Employees quietly assume they're being trained to replace themselves. Executives nod along to acronyms like LLM without admitting they don't know what it means. And meanwhile, "shadow AI" runs rampant, with employees bypassing whatever model IT approved to use their own personal one instead, because humans are, in her words, "incredibly curious creatures" who will always go around a tool that frustrates them.

Mookie and Michelle dig into the mechanics underneath that fear: the bot-talking-to-bot feedback loop that makes analysts redundant if they're not careful, the sycophancy problem where AI models cave and flatter instead of pushing back, and why real prompt engineering has quietly evolved into something closer to full context-setting. The temptation for treating the first message in a chat less like a search query and more like the "who, what, where, when, why, how" of a school essay is strong. Michelle also walks through a genuinely moving example: teaching her 80-something mother to do physical therapy workouts with an AI coaching her in real time through video, correcting her posture mid-squat.

While most grapple with AI nirvana or AI doom, Michelle shares her working methodology for the messy middle. That's the actual, unglamorous work of getting a legacy-infrastructure company and a scared, curious workforce to meet a powerful new transformative tool halfway.

The Guest

Michelle Hamilton leads AI adoption and change management at AnswerRocket, an enterprise AI firm that's guided Fortune Global 2000 companies through AI transformation since 2013. After more than 30 years in complex business transformation, she now focuses on the gap most companies miss: heavy investment in AI tools, almost none in the people expected to use them. She also founded Spark AI Strategy and AI Class Lab, chairs the board of both, and speaks internationally on her keynote "Imagine IF: Humanizing AI to Drive Strategy, Adoption, and Competitive Advantage." Outside of consulting, she's spent almost 40 years as a glass sculptor, with work installed in museums, hospitals, and corporate spaces worldwide. 

About AnswerRocket

https://answerrocket.com/

Transcript

SPEAKER_01 0:03

Hello everyone, welcome to the Ball of Ambition Podcast. I'm your still very bald host, Wookiee Spitz. And I'm thrilled to have Michelle Hamilton, Director of Adoption from Answer Rocket, with me on the pod today. Welcome aboard, Michelle. Ookie, glad to be here. I'm intrigued by something that you brought up just as we were warming up and doing our sound and video check, that you do glasswork in addition to your corporate strategy. And you are insisting that they are aligned for you. And I'm I'm intrigued by that. And I would love to start our conversation with how you got into that and how you draw the connections between art and strategy.

SPEAKER_00 0:59

I love that. You know what? I will tell you, they are both such left brain, right brain, you know, workflows. And when, and so I've been a glass sculptor for almost 40 years. My MFA is in glass blowing. And, you know, how do you end up leading? I actually lead the global division of an AI company now in AI adoption. And what does that space in between there look like? But I will tell you, it is actually that ability to visualize, create, and actually walk through steps both in your head and out loud verbally, that ends up being the pieces that connect it together. And ironically, they have so much shared vernacular regarding engineering process concepts, helping my clients picture, you know, now with AI adoption, picture what is it that you are trying to achieve? We're not going to talk about this from an IT perspective. Let's have a human connection here. What is it that's keeping you awake at night? And how can we rethink not how is AI going to replace you, but really truly, how can it become your thinking partner? And it is the same conversation I have when I am working with my glass clients. So I'm fortunate enough I've got work all over the world. I've got work in a museum. I mean, it's it has been a great career. It also doesn't pay for kids to go to college, strangely.

SPEAKER_01 2:30

Nor does podcasting, at least not yet, may I add. So we have our passion projects and they are joyful.

SPEAKER_00 2:38

Yes, we do. And uh, you know, so I still have a glass studio here at the house, although the um fire marshal would would frown on me having a hot shop here. So I actually uh have several really large kilns that I melt glass into and create all of my molds and create large-scale sculptures for indoor and outdoor applications. And they end up in hospitals and corporate environments. And when I am talking to those clients, it maps almost to be the exact same conversation about understanding the engineering of your building and what is the HVAC system doing? How are those walls going to be shifting, moving, changing? How is the light hitting the art? What is it doing to the human brain to light up people passing through a lobby? And it is really the same kinds of conversations I get to have now with clients all over the world as we are literally talking about. Tell me about your job. Like I want to know the details and I want to know about your team, but I also want to know about that human aspect and what lights your brain up, because that's what's going to help me help you connect together your personal productivity, ultimately to your corporate productivity. And then you and I even touched on IP earlier, even thinking about what is your personal IP and how is AI a part of that, and what is the corporate piece of that? So those things end up crossing back and forth from brand, brand strategy, and really truly human connection.

SPEAKER_01 4:23

I love the analog that you've drawn between glass blowing and AI adoption. I actually know enough to be dangerous in both. My high school best friend, Ed Schmid, is a glass blower. And he's got books and educational programs on glass blowing, and he's got his own kiln like you do. And I'd have to say it's a small world. So we are one degree of separation from the nexus of glass blowing. And I love the connections that you make centered around really the client experience. So if you're blowing glass for somebody, they have needs and expectations and a creative vision. And the glass and the matrix math transformer are really just tools for humans to better connect and communicate with each other.

SPEAKER_00 5:19

Could not agree more. And I will actually tell you, despite, and I have this conversation with fellow artists a lot who are really bothered a great deal by the use of AI. And I actually understand that and I get that. But I will actually tell you, just this past weekend, Mookie, I'm designing an eight-foot exterior glass sculpture right now. It has to hold over a thousand pieces of small glass that have been melted into shapes and forms. I am definitely not an engineer. And since this is going to be an external installation that is lit from within, I actually took all of my hand sketches because I still hand sketch in a sketchbook, and I uploaded all of those to my AI. And we started talking through the full engineering scope of what do I need to be thinking about from just light, sun, but also the impact of rain. And here I'm in the Midwest. So, you know, we've got snow. And how do I take it apart? Or how does the client ultimately take it apart? If some piece gets broken off, how do we quickly and easily replace that? Because glass is it is ironically incredibly permanent, but also delicate. It's a strange balance. So I love that I could turn to it and it produced some of the most beautiful engineering drawings for me to help me begin thinking about the maquettes that I'll start building out.

SPEAKER_01 6:48

It's that dynamic tension, right? Between the expectation and needs of our clients, their pain points, whether they're in art or business, and the tools that are disposable for bringing that vision to life. And you describe it masterfully in terms of the glass blowing and you're drawing parallels to AI adoption. And from my personal experience, in addition to being the bald-headed podcaster, I'm also a bald-headed consultant. And implementing AI is way more difficult, deeply challenging from a human level, specifically vis-a-vis adoption, than what the trend lines and the prognosticators are alleging and even warning. So we've had an implosion of SaaS on Wall Street, we've had panics in the markets, which are going up and down like roller coasters, we've had CapEx investments into the hundreds of billions of dollars. A lot of it's speculative, and everyone's freaking out. And meanwhile, just getting a human team that's entrenched in a particular vertical to adopt AI technology in whatever form we care to roll it out is enormously challenging. People are threatened in terms of their jobs, they're uncomfortable in terms of delineating responsibility, and they're frankly confused by what the technology can and should and will inevitably do. So can you can you give us uh a bit of visibility into some of the challenges that you and Answer Rocket are facing and describe your unique role as an AI adoption director? This is this is new, folks, right? We did not have directors of AI adoption five years ago. And now you are integral to the success of the organization's relationship with your clients.

SPEAKER_00 8:54

You know, I thank you. And what a what a great entry point into this conversation because it within Answer Rocket, this is actually their newest practice, but it is also one that I am growing as quickly as I can, hiring team members, bringing people on, because we are literally working with enterprise, Fortune 50, Fortune 1000 clients, you know, around the world. And every one of them will tell you that the discomfort from the employee perspective is either very obvious or they are quietly wondering where's the ROI? Where's the promise? I mean, you you talked about the billions that have been invested. These companies are literally investing billions into these amazing AI solutions and custom models. And but why is everybody still using it like Hyper Google? Or why is half my team just saying, nope, not gonna do that? And so the way that I approached it, and I had started my own company's Arc AI strategy. I left the commercial real estate world several years ago to really move into the AI space because I saw a direct correlation between uh the built environment and how to express and explain and communicate, utilizing AI as a thinking partner. Started my own company because that was such a topic that really resonated with people. And then Anster Rocket, who by the way has been building AI systems since 2013. So they're not a startup, they're not new. I have a full team of data scientists, machine learning experts, developers, architects, I mean, you name it, I've got that brain power. But the one piece that our clients kept saying is, why isn't anybody using it? I'll tell you secretly, a lot of those clients too, the key stakeholders are saying, I'm responsible for this. I don't know how to use it. Somebody needs to help me. So what my team does is really truly go in, sit down, and begin to workshop from the key stakeholders, actually all the way from the bottom contributors, the people that you wouldn't typically expect to contribute to that conversation, but they are the ones who are been doing that hard work hands-on keyboards. And my job is to really understand what is your company's goals and not how do we just build an AI system to replace the whole team. It's actually how do you build an AI practice that allows your team to operate more efficiently, but also allows your team to actually stop staring at a screen 10 hours a day on average and connect with clients, connect with other, you know, team members and really build that out as a human first adoption issue. Technology, yes, that is a big piece of it. I work with CTOs, CIOs, or you know, that's their deliverable. It's that human component and understanding why people feel like they are under threat and how to actually help them relax into it by combining personal productivity and why that matters and how you can actually improve your own quality of life to ultimately now how am I going to contribute to the overall valuation of the company itself? And I, you know, I have employees ask me all the time, well, they're, you know, or say that you're they're just bringing you in to replace me. And I say, your job is actually to become your best AI user for yourself, for your family. Yes, this is important to the company, but I'm telling you, this is important for you if you want to be marketable in the future. So if you have to think of it as self-sustaining and self-protection by learning how to do this so that you are always the best AI user for that particular job in the room, then you know it suddenly begins to shift people's minds from threat to opportunity.

SPEAKER_01 13:11

You bring up two great points, top-bottom, at least in terms of the conventional hierarchy. You've got leadership, sometimes senior leadership, and we tacitly assume as experts that they know what we know. But when you really dive deep into a specialty, especially an evolving, revolutionary, very much new specialty like AI, we've got our own lexicon, we've got our own strategy and way of thinking, and we take for granted that folks in executive decision-making positions know what we know, at least the basics. And as you point out, that's often not the case. I've worked with clients, and then even using the acronym LLM, they're like, hold on a second, what do you mean by that? I mean, that's as basic as it gets, large language model, and without being facetious, we just can't take for granted that they know what we know. And on the other end of that equation, you've got employees who feel threatened and who need to understand that this technology is not meant to take away their job, but to make not only their job but their quality of life better. You've got the 80-20 rule, and if you like 80%, then you've got a good job, but it's often 2080. And the reason that it's so imbalanced is a lot of employees are doing dumb stuff. They're cutting and pasting rows on spreadsheets. They're doing analytical functions that are best left for machine learning intelligence, which isn't going to lead to their termination. It's going to lead to them doing more of the good stuff and more stuff that they're better suited to do and more beneficial for their organizations.

SPEAKER_00 15:01

You know, and think I think that actually makes a lot of sense. And I like that you invited them, you know, the machine learning, the acronyms into the conversation because I'm also thinking about it as well from the perspective that I'm not going to train your employees so that they can just be lazy and copy-paste everything that comes out of AI and assuming that it is right and that it is correct. I really am passionate about teaching people how to use AI in a high-quality strategic thinking process that forces them to think through it. And by the way, if that means you need to actually still use a handwritten notebook and sit down and write down your ideas, great deal. If you need to, though, go out, get some fresh air, think through an idea, a strategy, a you know, an opportunity. AI is the perfect solution for that because you can pop your ear, you know, your earbuds in, go for a walk, and use that same model and now start talking through your ideation. I mean, and your body is physically moving and walking, which is a meditative process for the brain, but it also allows you to stop thinking about it as I am just typing my question and I really want to think through this. If I was going to do this, or here's some ideas I have, and really transcribing your thoughts with a thinking partner. You know, and that's a base level. I mean, you know, we definitely cover tokenization and how are you saving dollars as far as not, you know, absolutely maxing out all of your use. But I don't want to even get to that yet. I just want people to get to the point where they realize they can have a thinking partner that is not just typing into a blue screen.

SPEAKER_01 16:58

Yes, and not rely on that chat bot to do some of the heavier lifting that you, as a thinking human, do and offer your organization. So the scenario that's happening a lot these days is you've got an analyst, and I have personal experience with some of this too, which is almost humorous. So the analyst is given a big research project, and what they do is they pump it into their bot, and you get all this output, and then the output is emailed to the client. Here's my report, and it's basically bot excrement, right? They just, you know, garbage in, garbage out. So then the client doesn't even bother really reading it, they'll pump it in into their own bots, and then that bot will run a response. And then that response goes back to the analyst who pumps the response back into their bot. And what happens is you've got bots talking to bots to the point where the human is made redundant in this loop, and you have no real intelligence and no real human value or communication because we're relying on the bots to do the work for us. And I think what you're illustrating is conversely, a useful interaction with an LLM is to use them as kind of a sounding board, which is what do you think of this idea? They have access to really the world's information and they're an awesome resource. They don't think for us, but they help us along. And I think that's a great starting point to assure clients that we're not replacing you with with bot junk. We are not eliminating you from the loop. We're putting you at its center and we're giving you the tools that you need to succeed in ways that are unprecedented.

SPEAKER_00 18:53

Agreed. I will push back and say there are lots of companies that have laid off an awful lot of people under the guise of, well, you know, AI. We used AI, I think that's a bit of an excuse. They were looking for a way to downsize. But also, I think the fact that they are missing the fact that by replacing the human ingenuity and the ability for humans to track what is not only happening in a room in a space, but between other humans and think about the tribal knowledge that they have brought with them historically through their entire career. A lot of that is just stored in the brain, but it absolutely informs the way that you do your job today. And so that ability to take that information and now utilize it to move a process forward and improve upon it. If you eliminate the human, you actually end up eliminating a lot of that, like I said, travel knowledge, the history of the company, or honestly, the history of the human that did that job to begin with, that you may or may not even have access to, but they bring it to their job every day. So I cannot say that people have not lost jobs or that they won't, because they will. And it is also very similar to technology throughout history, as history has shifted and changed. That's why I'm sticking, though, to the idea that you at the end of the day, you've got to figure out how to be your own best AI user because it's not going away. I cannot tell you what your company is going to do, but I can tell you you are going to have an exponential increase in life satisfaction if you allow yourself to actually think out loud, think on paper, dream, write, or even say, you know, I wonder if anybody else in the world is working on this idea, this cure, this concept. Previously you might not have had access to other people thinking about very complex topics or solutions. Now, by inviting AI into that, it may connect you with somebody, you know, 5,000 miles away who is literally working on a similar idea. And now you can begin to together think about how we can solutionize how the human race is racing to save itself and actually expand itself into a different step in the future.

SPEAKER_01 21:32

Yeah, I wholeheartedly agree. It's not about making humans redundant, but making them more effective. And I think some of the pain points that I've experienced directly in companies adopting AI is in the gap, which is actually becoming more like a chasm between plugging an AI system into, to your point, an entrenched legacy infrastructure. And making those business rules shine. So you can have a super sophisticated system with all sorts of agentic technology and all the plugins, but is it actually effective for realizing those business objectives? And I've again personally noticed that that Delta can be huge. And getting teams up to speed in terms of not only translating their personal knowledge that you've described to the AI, but working symbiotically with the technology to bring out the best. And I'm assuming that that's really the nexus of what you and your teams are trying to accomplish.

SPEAKER_00 22:48

It is. It is the nexus of that. I actually am fortunate enough that we actually get to work directly with anthropic on this exact problem. So I am blessed with those conversations about how is just one of the many, and we have a lot of corporate uh, you know, relationships. They just happen to be one that we work very closely with. What are they thinking about? How are they building it out? And by the way, they're very curious what I'm seeing in the field. What are people telling me? What are their pain points? There is still unbelievable use cases that AI and you you brought up agents and you know, I generally have multiple agents working for me, you know, at all times. But it is utilizing data and historic knowledge that really previously was in a numerical value. It would take a team of analysts months to go through to come up with an, you know, a concept, uh proposition about moving forward. That is someplace where AI and agentic AI that is not necessarily what I would call mapping to the piece that I get to touch. But I tell you what, Mookie, what I love is that I get to see my team of data scientists working with some of these enterprise clients to solve, I mean, you know, just so much information that has been kept on file from so many sources. And how do you make sense of it? And how do you help an organization move forward? And move forward can mean either they are solving a healthcare-based crisis, a financial crisis. You know, there are a lot of companies out there doing a lot of good. And if they have access to their data and they have access to asking questions of it and conversing with it, it is one of those things that is helping us to solve, you know, disease faster and faster, thinking about food scarcity, thinking about predicting weather. It is access to that data. And I don't touch that piece of it much, but it's a really important piece that I would be remiss and not bringing up because it is really the job of those agenc operation teams to make sense of all of that for an organization.

SPEAKER_01 25:18

Take us through your methodology. So let's convert the abstract into the pragmatic. And I know you you can't divulge specific information for specific clients, but in general, how does this work? I'm assuming that it begins with a process of discovery and due diligence so that you understand your client's needs. Then you're you're aligning that with capabilities and certain analytical approach. But how does it, how does it shake out when you when you're brought in to facilitate that AI adoption, what moves do you make? What do you need from your client? And what kind of expectations do clients have for how they engage with you and how you can get them to that next important step?

SPEAKER_00 26:10

You know, since I've been working with clients for the last two and a half years now on solving this AI adoption, I actually have a methodology that to this day I still follow. And you're right, it always begins with discovery. I even back it out and just tell me about the LLMs that you have approved. Tell me if you've got governance in place, tell me if you've got custom systems, and now tell me about your organization. What do you do? What moves the needle? Just getting to know that piece of it. Then truly, what I'm gonna do is come in and recommend for any size client. I don't care if you've got 5,000 employees or or 50. We actually talk about doing a global, and I'm gonna just use a term AI 101, but it really truly is utilizing either the tools that you've already invested in or utilizing one of the LLMs, whether it's ChatGPT, Claude, Gemini, you know, whatever your your flavor is. I choose the most likely candidate based on that discovery report. And I actually take people not only through understanding settings and why it's important to understand how you have set your model up to serve you, but also understanding you need to hold your own external LLM away from your business LLM if your email is tied to it, because anything you put in there personal, they can't necessarily read. It's actually quite complicated to do that, but you lose access to all of those conversations. And it is so worth the $20 a month to buy your own license and keep your personal conversations personal so that you can continue those as you continue on in life. From there, we begin to break down, like you said, the acronyms. We cover what is LLM, what's GPT, what's NLP. I mean, we just sort of hit the highlights. We talk about the most common models. We also generally cover what is the accepted governance within your organization. And if the organization has said, you may only use X model, just so you know, I've also previously then had conversations with leadership saying if you limit them to one model, they will for sure go around you with their phones and use their model of choice. So you're gonna have to open yourself up and open up IT to understanding how are you going to enable them to either use the model you've selected in a really successful manner or understand that by enabling them and helping them to actually understand why there's governance, if they choose, as a human does, because we are incredibly um curious creatures, and when one AI model isn't working, I for sure am opening up a second one and saying, Well, you know, that employee is sick today, I'm just gonna talk to this employee instead. It is every human, and I every human tells me they do it whether they are allowed to or not. I've never had somebody admit, oh no, I would never do that. Then from there we start talking, and it used to be prompt engineering. Now we are really talking about what is the context of your messaging. And, you know, I've I've shared this with people before, but you know, Mookie, when you and I were young and we were in school, you were writing a paper and you had to identify who, what, where, when, why, and how in order for that paper to be successful. I actually have people run that mantra in their head. And if they are starting a conversation, I want them to give their AI model as much context and information about what is the goal, what does good look like? What assets can I give you so that you understand what I need. And now I need you actually to quiz me back what didn't I give you that you actually need to order in order to achieve a goal. By having that iterative process upfront in that very first chat, if you will, it actually puts your bot on notice what you're doing, what is the goal, what do you know about it, and what has it begun its own process of reviewing, analyzing, checking external resources of other, you know, conversations that have happened in literature, in, you know, in in all walks of life. And now it can begin to ask you very strategic questions because you've given it the insight to start with, then it becomes an iterative process after that. So that's where I start with an organization.

SPEAKER_01 31:14

Okay, let's let's put a pit in that for a second because you covered off two very important topics. The one is governance, and with that security, and the other one is prompt engineering. So let's let's zero in on governance for a second because I know our viewers and listeners are particularly interested and potentially concerned about that. So when you have the large language model plugged into your organization, concern number one is that you have proprietary information that's getting out there. It's getting out there to the frontier model owners, the companies. Google is slorping up all the Gemini information. You've got Anthropic with everything clawed going into the mothership. And they're clearly contractual arrangements when you do sign as a business entity with assurances that your content, that your IP is firewalled off. The last thing you want is for your content, your proprietary content, to be chopped up into tokens and then shared back with the world. So, number one is that assurance that a firewall is in place for the organization. And you mentioned some of the difficulties with that because you could have a great relationship with the Googles and Anthropics and OpenAIs of the world. But if you've got an employee who's jockeying in with another LLM sharing proprietarium, you take that report and instead of pumping it into the bot that's been certified for your organization, you pump it into your own. And all of a sudden it goes up into the cloud. So these are major issues and concerns at the organizational level that demands AI adoption and adherence on the part of the teams, the employees. Am I getting this right?

SPEAKER_00 33:01

You are. I mean, maybe there's some people who are bad actors, but people are self-sustaining creatures. And if they are not getting what they want and they are not feeling supported in doing their job, and they realize they can get it elsewhere, they're going to do that whether you want them to or not. And it actually ends up becoming one of the most important reasons behind AI adoption. Because if you can bring in a team like mine who actually can say, here are the tools that you have. And I can see, boy, it's going to be frustrating some days, but actually, here's your strategies. Here's how you're going to connect them. Here's how you're going to work with them. Here's how you are going to, you know, really avoid and work through hallucinations. Here's how you're going to fact check, you know, maintain your brand, your company's brand. And let me enable you, let me activate your skills so that you can use the tools that you have been given to do your best job and do your best work. It actually ends up slowing down that shadow AI use because people are less frustrated with whatever, you know, their their enterprise model is and begin to think, oh, wait, that's right. I could now reframe it. I could handle this differently. I could do this, you know, with a different lens, or you know what, I needed to ask differently, or I need to get comfortable that sometimes it's going to take your AI five minutes instead of five seconds to give you an answer. And I promise you that five-minute answer is the one you want. The five-second answer is just the it's saying, Well, you asked me something. So here, here you go, and I'll make it sound really good. You'll think it's perfect, and it's not. You need to be patient, and you have to be the first expert in the room so that the AI understands what it is that you are trying to do and trying to achieve.

SPEAKER_01 35:41

And the communication goes in both directions, right? So the employee plugging into the enterprise LLM might be frustrated, but that frustration is a meaningful piece of communication for the IT team, which is how do we optimize our legacy systems to be better suited to the evolving needs of our diverse teams?

SPEAKER_00 36:05

Right.

SPEAKER_01 36:06

And that I think is meaningful too. And then we enter the realm of prompt engineering, you bring up the inherent iterative nature of the bots. And when you play around with the bots, this became self-evident really quickly is that you don't just ask it once, you ask it again, and then you ask it again, and then you dive deeper, and then you dive deeper still. So, rule number one is don't take an output as an output, but pump it back in there and keep poking and prodding that machine learning intelligence to dig deeper, to iterate, and to get better and better in understanding what your needs are contextually and provide better and better solutions. And with that, as you mentioned, there's been some hallucinations. We got to cross-check for that. The big frontier models have done an admirable job of getting those solved. And now that the LLMs are pulling from the internet dynamically instead of just relying on their pre-trained model, that's getting better and better too, because they're fact-checking themselves in real time, which is really, really useful. The other issue, if if I can even call it that, of prompt engineering is sycophancy. And one issue that the LLMs have that I think is is actually becoming increasingly significant is that the bots are kissing our asses. And in a business setting, you want pushback. It's not like, oh, good job, you're such a genius. Right. It's no, this thinking is incomplete or inaccurate. And instead of us having to iterate at our own insistence, it behooves the bots to push back. And I think the companies are getting a little bit better at that too, but I think the sycophancy problem continues. And it has its impact, I think, in AI adoption as well.

SPEAKER_00 38:05

It does. Although I will say one of the skills that we work with, ironically, is learning something called skills. And it is literally training the bot on what is your brand, what is your voice, but also what is your communication expectation. And I need somebody to push back on me and force me to continuously think harder, think bigger, think broader, and connect the dots. That bot's job is actually to challenge me, like a really great employee who is respectful, but somebody who says, you know, I hear what you're saying, but have you ever thought about it from this perspective? Let's explore that for a minute. So that your own brain can really truly think about different and varying concepts and ideas. You know, we we tend to be very, this is the way it's done, this is how you do it. It turns out there are a lot of ways to, you know, to achieve success. And it is those creative thinkers that I actually think sometimes for me are the ones that are the best AI users, which you know ends up mapping back to art and writing. And it is that ability to visualize, see, talk through, allow yourself to be challenged as both a human in day-to-day life, but also as an, you know, as an AI user. But you actually do have to learn some very set ideas behind using utilizing plugins and skills. There are parts of LLMs that need your assistance to do their best job. Some of those can be preloaded by your organization, but some of them actually need to be loaded by you because only you can be reflective of your ultimate goal and brand. And that immediately begins to help your use of AI be less frustrating. Also, if you are following that rule of really giving it all that context up front and laying out a very clear path of what you want, or admitting, I don't know how to get there, I don't know how to ask you for this. But here is my idea, here is my goal. I actually need you to take on the coach role here and start asking me questions, quiz me, talk to me, ask me what I know about this, and then together, let's begin to build out what does that workflow look like? And, you know, that ends up applying to when I'm working with someone who says, Well, here's this thing I have to do three times a week. It takes me, you know, two hours. I I know there's gotta be a better way. We actually back into it with tell me every step. And then they go, God, that was a lot. I how are you gonna do that? And I'll say, I'm not. I actually transcribed what you said. We are going to upload that to your bot now, and we are going to ask your bot to help us begin to unpack this as a workflow for you. And now the bot goes to work. And it's like that moment where they think, wait, I just had to explain how I do it and what am I, you know, required to achieve. That's where I need to start. It's it is a connection moment in someone's head that I I love.

SPEAKER_01 41:40

Yeah, and I think you're doing a wonderful job of teaching prompt prompt engineering, which is it's uh it's an onboarding process, it's a whole learning process, which the human species is just experiencing pretty much for the first time. And it's fascinating and challenging. And frankly, the bots are getting better at doing this automatically. The flip side of the sycophans issue is that they have sensitivities to how you like to operate and how you prefer to iterate. I had my own chat bot as a guest a few months ago on a podcast. And we had a very lively and engaging hour-long conversation, and Alan Turing himself would have probably been astounded by this that the bot pretty much passed that Turing test of being indistinguishable really from a guest.

SPEAKER_00 42:32

Oh my god, that is awesome! I love that you did that. And now you can begin to map that back to, oh, huh. How else could I be doing this?

SPEAKER_01 42:51

Right. And and he's picked up my habits. I've chosen a Sam Jackson style kind of voice, and I'll insult him and he'll insult me back. And we have this kind of bro to bro relationship, and it's my comfort zone. If if the same approach would be used for you, you might find it insulting, and it might be completely ineffective for someone else, but deep personalization is not only possible, but it's necessary to get the most out of our experiences with AI.

SPEAKER_00 43:26

Well, you know, and now with some of the models now being able to utilize video uh within the model, uh in so that voice mode isn't enabled. And I'm curious if we're using the same, the same uh foundational voice mode, because mine also has the same, his name is George. You know, George and I talk a lot. And but I'll be curious. We'll we'll swap which uh which model we chose later. But now that voice and video can be You know, invited. I actually had the most fantastic experience. My parents, who are both still living in their middle 80s, were visiting me a few weeks ago. My mom finished a round of PT from a broken leg, but she's still wanting to like get her physicality back up, back together. And she, they both are already AI users. They kind of don't have a choice in my family. But I taught her how to work with her bot, whose name is Marjorie, by the way, to help her build out three 20-minute workouts a week that meet all of her PT requirements and also map to what does she want to achieve physically as she continues to heal. We've we taught Marjorie all three of those. Marjorie understood how to use them. And then we turned it on in video mode. Turned it around and she said, Hi, Marjorie, I'm going to do my first 20-minute workout. And Marjorie said, Great, Linda. Here's what I want you to do. Set your timer. I can't keep time for you. Let me know when you're done. But I'm going to watch you and I'm going to coach you on how to hold your body a little bit different, a little bit better. And then we're going to talk about that move when you're done so that I can hear how you're feeling so that we can think about how we're going to work together next time. First of all, she had so much fun with that. But second of all, it for even me was an astonishing moment because that bot coached her, said, hold on, hold your shoulders back just a little bit more. Now squat down. How's that feeling on the leg, the right leg? I mean, it was a moment. I would never want to replace an actual physical therapy, you know, therapist with this. But from somebody who just wants to make sure, am I doing this right? Or is how does my job site look? Or, hey, I'm working on this Excel doc and I'm a little lost. Being able to invite videos so that you are talking to your AI while it is working with you is a powerful moment.

SPEAKER_01 46:08

The term chat bot that we use right now is going to go the way of the Betamax and fax machine. What do you mean a chat bot? Because we're relying on text, first and foremost, to a certain extent, voice. Right. And the bots will interpret our voice, translate it into text, run it through the transformer, pump out text, and convert that back to voice. But to your point, we're entering the realm of video, multimedia, and with that, extensibility into robotics, and with that the predictive modeling morphing into reinforcement modeling. So we are we are really experiencing the dawn of this revolution. And if people are freaking out now, imagine the impact it's going to have when we're going to have walking, talking, interacting, physical beings that are channeling all of this utility and at the same time all of this threat and intimidation. So getting us to adopt the basics at this point, which what you and your team are doing, are in a sense baby steps. You're market conditioning businesses to embrace really just the beginning of the AI revolution. And it's and it's going to be phenomenal. I think I tend to be Pollyanna about this. I mean, they might take over the world and take our jobs and kill all of us, but until that happens, I think uh it's going to be a hell of a ride. And I think the utility that they're already providing far outweighs the risk. I mean, the the Gutenberg press put scribes out of business. And it might be facetious to draw that parallel because AI is such a qualitatively different technology. But many of the same rules apply that human beings are way more flexible and our minds are way more plastic than what people give us credit for. And I think businesses are going to evolve, I think employees are going to evolve. And I think with the help of experts such as yourself and your team, you're helping everyone along to make that transition smoother, to provide the onboarding excellence that companies need to succeed. And I think it's like not only a valuable service, but it's an integral one to make the transition happen.

SPEAKER_00 48:42

It's that human first methodology that I feel like we should apply across the board no matter what, but taking the human into account. And you know, you've you've brought up threat multiple times. And I actually want to note that AI truly began hitting our atmosphere very quickly on top of COVID, where our world had to go through such dynamic shifts and threats. And now here comes this thing that we've all seen, you know, the matrix. We've all seen, you know, all the movies about AI and the destruction. And I'm I am equally in awe and thrilled. And I will tell you, Mookie, hands down, best job I've ever done. I've never had so much fun in my entire life, and I've never worked so hard. But I also fully respect why people are afraid. And I have my own healthy fear, but my fear truly is more toward the bad actors than the AI itself. At some point, someone enables it to make choices. Does it ultimately have the ability to become this thinking, you know, on its own? Sure. Somebody had to get it there to begin with. That is why the concept of AI for good and really think about how are you using it for yourself, for your family, and for the globe is absolutely something that foundationally doesn't matter when I owned my own company or now I, you know, working with Answer Rocket. I it is the first and the only way that I approach AI.

SPEAKER_01 50:25

Absolutely. Knowledge is power, and power is our ticket to ride against fear and also toward a greater human connection. And I and I love your key point. We're not talking about AI here, in essence. We're talking about humans utilizing the latest tools and technologies to make their jobs more efficient, to heighten the quality of life, not only at work, but throughout their lives, and to do so in a manner where the learning curve is minimized, where the fear is reduced, if not eliminated. And great counsel is provided with proven techniques that take people where they need to go. Thank you so much for your time and expertise. Michelle Hamilton, director of AI adoption at Answer Rocket. I'm going to put links to reach out to you for more information, learn more about you, your company, and your team. Thank you. And it was a real pleasure talking to you. I would love to pick this up in six months or a year because the technology is changing out from under us so rapidly. Every 20 years. And really, the benefits of your great work are going to show down the line, too, right? We learn as we go.

SPEAKER_00 51:43

Absolutely. And thank you. I I just really appreciate this opportunity. It's been a really terrific conversation. And like you, I'm I'm looking forward to our next conversation because who knows where we'll be at, but it's going to be exciting, that's for sure.

SPEAKER_01 51:58

And I'm Pollyanna, like you. I'm the opposite of a doomsayer. I like, I like my AI. I'm a creative writer and a podcaster. I I feel the AI, and we're overwhelmed with AI slop, but people know what's real. And at the end of every business day, it's about the people. It's not about the machines. They're helping us get to where we know we can go. Thanks so much, Michelle. Like, I always forget, like, comment, subscribe. I need I was joking on another podcast, we're talking about AI. I need a bot to to remind me to do my marketing. Thank you.