Latest / The Failure Gap / A Conversation With Josh Miramant, Founder/CEO, Blue Orange Digital

Transcript

A Conversation With Josh Miramant, Founder/CEO, Blue Orange Digital · The Failure Gap

• speaker-0 (00:00.162) Hello and welcome to the Failure Gap, where we talk with leaders about closing the space between agreement and alignment. We love talking with interesting people and today we're joined by Josh Miramant. Josh is the founder and CEO of Blue Orange Digital, a leading data and AI services firm. He works closely with executives and investment leaders to define strategies that harness AI and agent systems for business impact. His focus is on helping organizations move beyond experimentation and into production-ready deployments that drive measurable outcomes. Josh, welcome to the Failure Gap. speaker-1 (00:33.55) Thanks so much for having me, Julie. Excited to be here. speaker-0 (00:35.81) Yeah, I'm so looking forward to this conversation because I think it's a really rich area right now for organizations and people who are in the failure gap around AI. But before we get to that, I would love for you to just give our listeners a little bit of your backstory and how you came to be the founder and CEO of Blue Orange. speaker-1 (00:54.658) Yeah, thanks for, it's an exciting story. And I would say not the most, linear path to get here. So, actually have a background. I've always been in technology and living around that for all my work. But I, originally was kind of on a trajectory to be in politics. My family's very in politics and I came up working in political races when I was young through school. And like all, great accidental stories, the, one of the fundraisers at a campaign I was at. was in private equity and had this great idea to start a technology company. And I was either going off to law town, law school at Georgetown or going off to start this company. he ended up recruiting me in, and I was the first employee at a tech company that on day one kind of exploded. And so I was sitting my, instead of my post-graduate education being in law, it was in startups and scaling. And I was right in the room, both on building out technology, and working closely with fundraising. And just watching a business go from two of us in a row house in Georgetown to a couple hundred fifty people in a couple of years and then even growing beyond that and just being a part of that journey. And so it was really good hands on practice with getting into kind of a exposure and tech startup world. I then went over and really wanted to kind of kind of outscaled my skillset. The company jumped beyond what I was capable of offering and it was moving beyond me. And so. My young petulance said, I want to go do it all my own. So I went off and went off to San Francisco and started my second company in HR and talent analytics, which got me closer to data and analytics, fundraised a bunch of money and on selling that company. We effectively took the services work that we were doing out of that and turned it into the professional services arm today. Continuing down the track of working with analytics and helping companies kind of figure out how. to have better insights inside of their company. And I really loved that work. I also liked moving from traditional venture-backed SaaS, which gets a lot of the startup notoriety into companies that run really well with a P and L and are managed by their ability to sell and build good foundation and have a strong business model. And so I really took to that and that was about a decade ago. So we just turned our 10th year old in June this year. So it was really fun to be on that journey. So it's been a... speaker-1 (03:19.534) Not the most direct path. wasn't thinking this out of school, but kind of stumbled in by really good mentors and opportunities and a lot of, would say, I guess a lot of not knowing what I was getting to the right level of ignorance before each decision. here we are. straight. Ignorance is a very important entrepreneurial endeavor for sure. speaker-0 (03:41.642) I love that. Yeah, I would say for any of our listeners, I always like to remind you if you are trying to figure out where you're going next, or if you're even just getting started in your career, remember that there is no straight line to leadership. it sounds, Josh, like one of the things that you've done is you've said yes to opportunities when they presented themselves, even if they represented what might be considered a turn away from what you were aiming for. And I think that sense of curiosity and adventure serves you well. speaker-1 (04:12.27) couldn't agree more in like, first I've lived in practices and it's a developed skill. think I can, you can like summarize 20 years of career in a couple seconds. And it's like, man, just leaned into it. And then there's also the process that is the thing I love most about building business and then which I think I love the idea of the failure gap and identifying that the transom from kind of what's not working to how it become that works. that is. building a business that is making a decision to make a different career. Like that's everything. I mean, if you break it down to the catalysts of those decisions, getting good at that and having process around that action is really everything that holds you back from success. So even as you frame it, like how do you assess a decision and move from, you know, make a big leap, but doing it in a calculated way. And that's, that's a lot of where I have like these seminal moments, which I remember, I'm not exactly sure how I developed it, but it's been a practice my whole life. And even in building a business and growing and scaling and all these different things that you kind of assess, it's kind of an amalgam of a bunch of different frameworks and processes that kind of have it net out in the positive a lot. it's definitely being open to it, but also being thoughtful, you know, calculative and strategic against it as kind of those, the tensions that kind of live in this, the kind of diametric opposites in some ways, but they're both very important in balance. speaker-0 (05:38.646) Yeah, I think you have to hold the space for both of those things in yourself and also help the people who are coming along on the ride with you to understand that as well. That we can have big ambitions and be excited about them and we need rigor and discipline as we pursue them. speaker-1 (05:55.566) It's a really good point. There's a funny thing that I, and I find that there's actually a whole bunch of, you know, the failure gaps I've identified and just pulling back to that. But even just as a founder, one thing when I'm talking with other founders that I just see, I was writing this post when we turned 10 years old and I was going to do my, actually my, wonderful partner, Trinity. She was like, you should write a post or your 10 lessons for 10 years. I'm like, that's good clickbait. Like, yeah, okay. And I did that. But my other article was. the just endless paradoxes of founders, which is these things where you get really great advice, really smart people, and both things can be true. And there are so many of them. Like I came up in fundraising and it was like, my CEO has always raised as much money as you can get. And then you were like, well, you talk to early advisors and you're like, well, don't dilute your company. And you're like, okay. Or it's with product, it's like, you know, make sure you don't build, you know, to spec to a single client, but then make sure you incorporate as much user feedback. And it's like, Both are true, but they can literally lead to diametric opposites or, you know, it's like micromanagement versus delegate delegation layers, like the in the weeds versus hands off. And it's like, there's so many things that I would say you, you can get really apt and applicable, incredible advice that is, that have these diametric tensions. And I feel like this is another one where it's like, you know, be open and kind of take the leap, but also be very calculative and strategic. And it's, I feel. balancing between these contradictions, these paradoxes, I think there's so much of that advice that exists in business and in founding and really leads to lot of taking the wrong side of that assessment can lead to where you kind of get stuck in the failure as opposed to pushing through as well. speaker-0 (07:36.45) Yeah, yeah, or it feels intimidating or overwhelming or confusing. And I would say for our listeners, if you are a founder or if you are hoping to be a founder of a company, some of these things are really important to recognize that making, we always talk when we think about moving from agreement to alignment, that making things visible is so important. And there are no necessarily answers or solutions for some of the things, Josh, that you're talking about, but by making them visible and making those trade-offs or those tensions, more available to you and your team to work through together, you can move into alignment on how are we going to tackle this? How are we going to show up in this diametrically opposing space and be successful? And I think that is a founder mentality is you're constantly navigating tensions or trade-offs that feel like they are going in different directions or they require different things, but you have to be able to bring them together in a productive way. speaker-1 (08:34.734) Absolutely, as well. speaker-0 (08:36.363) Yeah, I would love Josh to get links to those articles or those posts and we can put them in the show notes for anybody who might be interested in taking a deeper dive into them. speaker-1 (08:45.23) Absolutely. Yeah, they're fun to write. was mostly, I, one of the favorite parts about the paradoxes is I, I'm in a founder group where there's about a thousand founders and they're all at different stages, but I kind of put it out in our Slack group and unbelievable torrent pouring in from a bunch of founders kind of giving their examples and man, there's some good ones. So I tried to, included that list in the article as well. So it definitely was cathartic as a founder to be like, I got that exact advice. also, there's the contradictory one that you got. And you're like, I got both of those and had to struggle with that same thing. felt very cathartic. speaker-0 (09:21.335) And they can both be true. speaker-1 (09:22.958) And that's exactly it. This is a challenge I always find. when I think, you know, as I really love the theme of the podcast, because it made me think about, you know, how do we find failure, move through it, is this is where I ran into a lot of them wasn't out of just wrong conviction or any of these things. It's like, it's maybe waiting, like either not knowing some of these areas or not taking a process to it's like they might be both right, but you have to take the context to it. So you can end up like. The amount of times I've been persuaded by a mentor with really good advocacy or a book that's really persuasive, I feel like, okay, this is the clearest path. And it's like, we live in the nuance that is the challenge of all of this. So it is absolutely both can be right. And it's just how you choose to navigate clarity. Like I love the clarity that you bring to these things is really critical. speaker-0 (10:12.002) Absolutely, and I think the founders face can be so, gosh, a lot of people say I would love to start a business, but when you get into the nuts and bolts of it, it is scary, it's hard work, you have to be all in, and you have to be really thoughtful and diligent around how you're organizing around it. And that's where I think a lot of people get stuck in the failure gap and don't know where to start. And I think what you're saying is get comfortable with that. It's okay to feel like you're getting lots of different kinds of advice. Take what works for you and leave the rest behind, right? Like see how you can navigate that for yourself. I'm not just looking for some kind of truth or some kind of clear path forward. speaker-1 (10:52.75) I love, this is feels like a little bit in everything for me. And, you know, I don't think it's actually just for founders. This is like, it's one thing I lens a lot of, have an identity around it, but like, even you're just talking about whether you want to leave a job or making a decision for really anything. There are many frameworks that exist on how you think through different decisions, but you're, but even before you're coming with a framework for decision-making, it's even just sitting down to understand, um, It's the, is it the Rumsfields matrix? I can't remember the name of the matrix, but you have like the known knowns, the, I'm sorry, this a little bit. And then like the other category, the functions is the unknown unknowns. And then you have the swap. There's the known unknowns and the unknown knowns. actually I love this. I use this so often in like, much do I understand a thing before I'm ready to make a decision? And so you can't know your unknown unknowns. That's the thing where you ask for experience and you have to step outside yourself. And that's like, the opportunity in somebody where you get somebody to experience or read a book, you know, you're no knowns. Those are just things we are aware of and understand it's this attention, but like the other two categories, you can actually do a lot of coloring in the map a bit on those. And then you will, it will help define by abstraction what you need to like the unknown unknowns become something that you need to just either do you feel conviction with what you know, or do you need to go find out more about it? So before you get to decision frameworks and net present value or all these different ways of thinking about really good decision areas. I do that exercise all the time. I do it with personal decisions. I do it with business. I do it strategy. And it's a thing of just before you're making a decision, how well do you feel you know it? And the exercise of trying to put that down and categorize them, even just lightly, feels like it kind of brings a lot more color and kind of like fills in the map of what you don't know, which is very nice. speaker-0 (12:43.01) Yeah, I love that example. we can, again, for our listeners, we'll throw a link to that in the show notes. But I think having that kind of framework to assess where am I and where do I need to go deeper? Where can I move forward? And like, find your way through some of these complex decisions or maybe things that we don't have as much confidence in because we're not experts in the space, it can really help us. And kind of brings me to another topic, which is, you know, we've talked a little bit about this idea of individuals embracing AI and being able to use it for themselves. And I think about that four square. And I think for a lot of people, feels intimidating that there are a lot of unknown unknowns. How do you find your way through that? I'm curious because you and I both work with a lot of leaders who are right now today realizing that they are now late adopters. They're behind the curve because they haven't taken advantage of the last few years to just learn and experiment and grow. speaker-1 (13:22.315) Huge amount speaker-0 (13:39.23) as this technology gained speed. And now it has gained a lot of speed very quickly in the last year. So I'm curious. I'm curious as you think, think that's, know, leaders at often a very senior level in an organization being in the failure gap around AI. When you think about moving from agreeing that it would be great if I knew more about AI or if I was doing more with it myself to actually getting aligned and getting that done, what has, what stood out for you that is keeping people in the failure gap? speaker-1 (14:09.014) Yeah, great question. And like, I empathize with those trying to understand and step into this space deeply. And that like, I think there's, you frame that really well. And there's two categories of like, and I think we're really looking at right now, AI is this advent of large language models and kind of wide distribution that we have as different forms of integrating and products we invest in our businesses, we can use these tools individually. And there's just so many ways of what at the core is a general purpose technology, similar to like the internet or mobile. There's not just a one way use, if it's all usage, it's a broad application of how these integrate. So I empathize with those trying to figure this out. Cause it's like not just, it's not as simple as a question is how do you use AI? And you identified, I think a way that I look at, at least, you know, there's a lot of ways to break this down. I think breaking it down into the requisite pieces is a big part, but like there's individual use. And then there's like, how do you shift your business to be ready to be susceptible to the change in the tooling and the enablement that comes with the productivity opportunities around AI. And both are quite different because moving an organization and change management and adaptation and investment, that's where we live as a business at Blue Orange. That is a different process than the other challenge that also exists is understanding or seeing the aha moments. And I'm watching. I don't like to have fear around this because everything's learnable. And I still think that though it's moving so quickly that things we are using around individual use around AI will change and will update. And so there's some part of that means in change is disruption. Like the way we're using it now will change. think there's actually, have a few, I just released a post at the beginning of the year around my kind of predictions for 26, which I'm happy to be proven wrong on, but there's definitely a number of opinions. But going back to your question, sorry, the individual ask is, think that this is a good time to get started figuring that out. And what I mean by that is there's this early adopter tier where the way I was working with AI at the start of 25 and end of 25 was quite different. And there are things that are, there will be a lot of change in adoption, but there are stuff that has becoming just a clear value tenants that come out of that. speaker-1 (16:26.038) And so I think that is, you know, I wouldn't feel too far behind, but I'd also put a set of urgency on top of it because as a few things have changed in technology and all this money flowing in, they are just moving quickly to becoming more and more value. And it's not that you're not keeping up with the technology. It's that a competitor or somebody else is doing your work or some of your job is going to get better with that. So I don't see that like AI is replacing jobs is the Jensen one, you know, Quote is somebody that knows AI is going to like fill that role. That's the thing that is there some compounding gap that's happening right now of building a learning framework. So I didn't answer how yet, but I would say I empathize with it. And I do think it's thinking about how you individually consume AI because these are open tools. And then also how you think as an organization are distinctly different. Those are actually very important. There's different approaches to how you think about what your decisioning is there. speaker-0 (17:19.534) You know what I've observed, especially with senior leaders, think deeper in the organization, it's actually I see more desire to lean in and to figure out like how to how to use AI. But at the very senior levels, I often see people say, well, you know, I don't even know what I could do with it. Right. Almost like a lack of curiosity around that. And yet they are the people who are making very big investment decisions. Right. Exactly right. so it feels like there is a gap there that requires some individual ownership and curiosity around what is possible. And you start learning that by individual application, and then you're able to better assess opportunities for the organization. So there's almost a pull through. So you're not making decisions about something that you just simply don't understand at even a user level. speaker-1 (18:11.168) And I think that's the part you know it so well. And so getting into a little bit of when I've worked directly with founders, I'm sorry, leaders of organizations. Again, this is where our company spends most of our time. We are a technical implementation shop, but our consulting value is high and having figuring out where workflows are identified and working with the leaders to both get them on board with what this looks like. And I do think, as you say, it starts with individual like curiosity and exploration. And I think that before, so You can just, one thing I see and it actually, even with people that like experiment with technology and are comfortable, you're still seeing a divide occur that a lot of them feel that it's just, they have the abundance, they kind of get it, but they're not seeing it yield. And so there's actually a decent amount. If you take chat GPT and you fired it up today, there's actually quite a bit you have to do to get it to be really good at it. a large amount of tasks. Even if you're pretty comfortable with technology and you're kind of clear on what you're trying to do, you need, there's this whole thing about like, what context have you exposed it to? How do you ask it? This prompting thing, like how do you incorporate it into a workflow or how you capture ideas appropriately? There's actually quite a bit of infrastructure around it. And so understanding, one of the things that is the chicken or the egg without, know, if you're doing this in isolation, is actually identifying like the breaking down. I always look at this even before with automation in general, which ultimately infrastructure you put in place is like, what's the anatomy of a task that you're trying to do? The pieces of that are always a really good starting point because there's some things that AI is fantastic at and something that is not. And so if you actually understand the kind of the core thing that, uh, LLMs are very good at this, like generative output and some layer of proximity for judgment, like that's what it's kind of like a proxy for. There's still a whole bunch of pieces around that. Like you have the data, the action that you do with it, the outcome of that action, how that feedback feeds back in, and human verification of that judgment. But the prediction or the generation of what that LLM is producing, that's what LLMs are great at. just able, or AI is very good at. So... speaker-1 (20:30.562) I think there's this argument to be made of actually under when you take before you even go and play with something. Ask is like an LLM or identifying a task and saying, which part of this is the LLM good at? And then what do we need to ensure is available? And that question, it sounds a little abstract, but it is a exercise that I'd worked with a lot of senior leaders on that maybe don't feel familiar. And then if you just show how it does the part of the task that it's good at, they're like, and you explain that it's so eye opening, like, that's the part it's good at. And, and then there's, so I think that's a big piece is like knowing where it's good and not, and kind of mapping a couple examples that are familiar building a report for an end of month meeting. Okay. So what is the good part of the LLM? What do you need? Like, what do you do as a person to get that? go read these things. I sit down and I write out this, and then I build a report in this form and I have a template and it's put into a BI tool or an Excel tool. And then I go talk on it. Well. Part of those is not elements aren't particularly good at. You can't replace aggregating all your data. You still need to provide that to an AI model. So, but you can identify that taking it in and getting to some output it's fantastic at. So it's when you've broke down the full life cycle of that leaders task and then you show where it's very good at that is just the cascading effect of like, okay, let's look at other stages and like, let's look at these areas where, you know, this is what does it need for understanding? And then you can. think about how to make that available. And that's where the tooling and the systems become really, like that's just the stuff that we understand and solve. Like, okay, what if we want to reduce the effort that you don't need to figure out how the data becomes available to the AI? You want that instantly and always available on all of your corpus? That's the types of problems that we say, okay, you have that need, this is how you can solve that most efficiently. And that's the enablement function. And that's a lot of where we live. But it is really just starting with understanding what it's good at and what it's not, and just showcasing those examples and then stack two or three and then that's the cascading effect. It kind of rolls from there. speaker-0 (22:38.786) I know one thing I've just been really encouraging our team to do is play with it and be willing to experiment because I would agree with you. The way that I was using some of these models at the beginning of 2025 is very different from how I'm using them today because I spent a year playing, not actually solving business problems, but no, I mean, to some extent, solving some business problems, but for the most part, just learning, right? And you learn best when you put your hands on the keyboard and actually mess around with something. So I would encourage people to, if you want to move through the failure gap as an individual, think of some of the things that Josh is sharing here around how do you understand really surgically, and you might need to ask for help with this, how do you understand really surgically what part of something an LLM might be really good at, not expected to be able to just be a magic wand, right? But it can actually break things down and do specific parts of what you do very well for you. And then... Play around a little bit with what that looks like and how you could leverage it in different ways. Don't expect it to be perfect out of the box. speaker-1 (23:43.66) Such good advice. I'll tell you that one of the things that the two paths of fast learning for me are exactly, you mentioned them both and I'll just flag them and I tell everyone in my company to do this. an advocate and I tell my clients and people we work with, but it's experimentation and collaboration. Those two are, you mentioned both and I mean that, but I have learned as much through just playing and messing with, which the stakes are quite low. You can fire something up and not have an output like, The output of an LLM is very commoditizable. You can throw it away and move on and you can kind of understand the process. But even quicker is been through, I actually have a very active group of technology leaders. They're actually all working technology around a whole bunch of firms, about 500 of them in a group I'm in. And there's just a perpetual like, I did this, here's my result. That has been the biggest uplift I've had in the last 18 months. And I'll even directly like the amount of tactics that we're bringing and techniques we're bringing into our clients. The corpus of examples I have through 500 companies are, you know, I'd say that it's a very active group because of the excitement of AI. I have seen and iterated and then I'll take those and tests and we'll identify what works in the use case. And I'll see somebody using something in a different for a SAS tool. And I'm like, Oh, this is perfect for some analytics solution. And, you know, it's like the connection of partially thinking and fitting. That is unequivocally been the fastest way to learn, which, you know, isn't specific to LLMs by any means. It's really, or AI tools, it's really just to all learning. It's like that collaborative, you know, it's the best way to feel like, like, uh, you know, not being vulnerable about coming up with bad mistakes and just being able to put that out there. Like I like that it's new because a lot of people, like I come up learning how to build software. There's this, uh, it is this. site that has now made a little bit redundant because of, LLMs have scraped it all in and make it very good, but it's called, stack overflow. it's where every, anyone in code goes to look at code and share as a repository, but it famously had this, if you didn't ask a smart question, you'd get ridiculed by the community. And it's, and it was actually an intimidating, I remember as a developer coming up, I'm looking and I was like, okay, I need to ask a question. I know it can be answered by this group, but. speaker-1 (26:02.05) Let's make sure I'm really going to spend a lot of time worrying about if this question is good or not. LLM's taking that all away. The amount of like dumb questions, smart questions again, there isn't really one, but like the idea of you could just ask it anything and it's light, sick of fancy. We'll be like, great, try it, Josh, try again. That just opens up to like reducing the barrier to, you know, what I think one of the thresholds to efficient learning is just not like not having a barrier to being vulnerable on your, what you don't know. I think that's a really critical part. It's like the perfect catalyst where you want to ask how to use an AI tool. It will tell you, you can ask a bad question and we'll give you feedback and help you learn. addition to it's, it's a wonderful, uh, it's a tool that is very self adaptive in that way. So the variant entry is incredibly low for learning. speaker-0 (26:50.028) Yeah. And you know, if you are someone who likes the, radical, you can train your LLM to talk to you that way. speaker-1 (26:55.726) Yes, reduce the sycopensis. speaker-0 (26:58.162) If you thrive on that, But I think one of the things I really want to emphasize here is, and I'm always telling people this, you can ask my team, I'm telling them this all the time, but find low stakes way to learn so that when high stakes decisions come up, you can make them as well as possible. And I do think that is something about moving through the failure gap. If you want to get out of just passive agreement that something's a good idea, go out and put it into play in low stakes environments so that when it gets hard, you're prepared. Don't wait until then to try and figure something out. speaker-1 (27:31.566) I think there's a, and this is moving a little bit more into the organizational environment that when we, and it's not about people that are still trying to figure it out. Cause like say that you come in, you start to see the value, but there's this, I studied economics growing up in school and there's this, I think it's the Jovens, it's a good part of Jovens paradox right now, but it's the idea that if you, make something cheaper instead of it, it causes it to be used more often. It's one of these things that have come up. It was around the steam engine where, that got really cheap, became, yeah, so that's been, and that whole thing. And so now the distribution of goods got cheaper. Some more goods were shipped. That was the original thing. And then the steam engine that's been thrown around a lot. And so you're seeing the question then becomes with something becoming cheap, like what AI is able to produce. speaker-0 (28:09.752) People were traveling. speaker-1 (28:26.326) What is, there's this other effect in the downward sloping demand curve that things related to that good become more valuable. So let me get the example of coffee. Good goes way down. Then the demand for cream and sugar goes up because the cost of it goes up because cheap coffee, more things you mixed with coffee. So there's a relation between other goods. And so I think we've been looking a little bit at a high level about where that is with AI coming into an organization. It's like, okay, say literally doing Googling or reading books or doing analysis took is gone down other related tasks to go up in value. And so there's this, there's these kind of four new bottlenecks. use this consulting, of course, like try to make things memorable. We use this acronym of card C A R D and the things that go up that become critical when you're implementing AI and organization. The first is clarity. And so if you're not clear on what like, see how I mentioned the example of like understanding the anatomy of the task. Like where can you apply this? That you need or even how to ask something to a model is really critical. That prompting is just clearly asking in the way a model wants to know. So clarity is really critical. And if you get that result, how does it make impact? That's one area that when we're identifying a really good workflow, kind of across all industries and all integration, clarity is a big thing. What do you need to get into it? Where does it apply? The A is This is an interesting one. Ambition is the A. So it's this idea that the thing that has been completely cheapened a lot is just the ability to produce. It's no longer like if you're producing a report or a contract or whatever, anything with linguistics, all these things that are able to produce that very quickly, but you should be very ambitious with what you're trying to do. You can produce instead of just one finalized product that takes forever. It is so easy to produce 50 things and then just choose the best and amalgam and iterate and all that side. So you should that that is the area of application that works very well. The bottleneck is, you applying this in enough, you know, make 50 shots at goal instead of one? Like that's a new thing that's coming out of it, but you need to change your thinking. It's not like, okay, one big, we can make one big proposal and it takes our whole team to pull it together. Well, cost of generating, you implement AI well. speaker-1 (30:47.278) Cost of proposals are a lot easier. You can get your IP down there. So how do you go do five, 10 shots that go all as opposed to one? The R is that's like this ancillary thing around AI's relationships, whether it be internal, external, the premium on that is shooting up. It is just the idea of, yes, it's great that you can produce a tokenizable thing very well, like something you can create content on, but everything still in business connects to relationships, whether. how you run your dev team, to how you make sure your product mapping, how you build sales and marketing, it's communication out, how you build relationships with clients, it's all relationship driven at its core. That has become the premium, that's going up in value with commoditization of some pieces of generative language. And then the last thing is distribution, D. The D is how do you get it out to something? If instead, if you have something very valuable, You stay in everyone else can produce a lot of things that very value. There's more things in market. So channels and visibility are critical. looking at like the value of AI inside of a company, have to, once it's adopting, you still have to really go down and look at these kinds of four areas that are the new kind of constraints on how to get AI to be successful. It's like the clarity of what you're looking to do with it and how it connects the ambition. That might've been squeezed in just to fit to the card acronym, but it's the idea of being very motivated to get their relationships and distribution. That's like a big thing of like looking at where process integrates. speaker-0 (32:14.22) Well, I love Josh that, you know, we're consultants and so we like to think in four squares or in acronyms. speaker-1 (32:20.17) Yeah, you see my training coming out here. speaker-0 (32:22.612) We'll make it work. I love this. Yeah. And I think clarity, ambition, relationships and distribution. When you think about that in the context of AI organizationally, I'm going to go back to the fact that if as an individual leader, you haven't been experimenting and playing and learning about these tools, it's going to be hard for you actually to get those four things going with an organizational decision around AI. So. speaker-1 (32:24.526) I'm cadena brain. speaker-0 (32:51.766) I do think there is some individual impetus to like get aligned around your own development in this space and use that so that organizationally, when you need to create clarity, have ambition, build the relationships and dial in the distribution, you're doing it from a grounded space. You don't have to be an expert. You really don't. You don't have to know how to code, but you should understand prompting. You should understand what it's like to... develop a relationship almost with an AI agent or to leverage these tools as a user because that will help you to make better decisions organizationally as a leader. speaker-1 (33:32.652) I couldn't agree more. And again, going back to that point where one of the things that AI tools are fantastic at is teaching how to use them and just asking. that thing that couldn't be the biggest unlock when somebody feels comfortable, you know, there's the cliche and, know, don't want to throw my parents into the bus, but of course I still have to help them with technical issues on things on their equipment. And it's amazing how that has them teaching them, you know, teach the man to fish scenario where they can go on to AI and it's so explicit and teaches at their level. You know, they, they put the poor son helping them with IT problems out of a job here. So it's really good at like meeting you at your level and bringing through. one other thing that I use, and I'm sorry, I'm a, I am a technical consultant, so I think in frameworks, but there is one more that I really find helps is, the other thing to take note as your, as companies are thinking about where to, where there's opportunity inside of AI tools. And I spend a lot of time in specificity, which is like, okay, here's supply chain opportunity, sales and marketing. That is something I'm always happy those conversations, but still as a framework that I think about is there are areas where AI is kind of generally very good at, and there are areas where it's generally not. And so if you should think about where your business sits in that category, and you'll have more or less opportunities in it. So there are kind of like three levels that I've seen that are really where we categorize, and they're very broad because everything, know, consultants speak in frameworks, here we go. But there is this idea that... You know what it's really strong in generation at its core. This is it's let's take the name of what the, this new technology that, really exploded as a large language model. That language is a big piece of it, without getting the weeds. They are the technical weeds. Like what they're doing is tokenizing some form of language, whether it be English or code, or it doesn't matter what it is, something that has, the relationship in language. and it's able to produce things in that format as well. So we have. Everyone's played with this stuff. So there's some format that are able to use documents. But anything that at any part of that, if you're looking for that layer one, that stuff is going to be very attackable with AI solutions could impact it. So that's where, if you look at the first department that gets a lot of opportunity is usually sales and marketing. It's because that is the core foundational component. You look at the second, another one that has a huge amount of disruption in its coding, the language is code. That's another one. So these areas, you know, speaker-1 (35:57.194) super dense proprietary like supply chains out there that work in third party systems. There are pieces of that that have optimization, but it's not the entire like structure of the parts that you're using aren't a language. So both English and code, you see two areas that attack the quickest, which is sales and marketing stuff, which is that and code. So that layer is something to identify. But the thing around this, the second layer that's on top of it is that there is actually quite a bit of value in, has a little bit less, you know, that's one of those other pieces is that judgment and accountability. There's a ton of stuff in business that you're dealing with, whether it be how you build process and all that. And the creation of that doc, the generation of the process or a strategy effort can be enhanced with it, but it still has judgment and it still has decisioning. And that's very, that's not the strength of LLMs. They can help with ideas, they can read docs. So you have a shared responsibility. So anything that has tokenizable condition is that first level. Second, have that judgment and accountability, it becomes a shared responsibility. And then when you move down to like, there's this thing in world of bits versus bytes, like when you're dealing with something physical, that's like the least attackable with the AI that we're using, large language models. So there will be advances in robotics and AI and when AI is broadly applied, but the real big advances in LMS, not very good at interacting with physical representation. That's not what they're good at. guessing. So that's like the least connecting. So it's kind of looking at what really operates in a language and then how much is that's the first level. it's entirely that it's very attackable and really good to put your energy into. If it has a lot of judgment and accountability into there's a shared responsibility, but it's less. And then if there's like physical structure around it, that's the least attackable. So it's not really great everywhere and you'd want to do it in your highest opportunity, lowest cost quadrant. slipping in another little framework over there for you. speaker-0 (37:53.804) Yeah. And I would say, know, luckily, if we go back to the individual, much of what individuals can play around with in a low stakes environment is very language driven. Right. How do you get help writing an email, writing a LinkedIn post, doing different things? Yeah. How do you start to experiment in those low stakes spaces? speaker-1 (38:04.163) Exactly. speaker-1 (38:12.942) And I'm excited too, there's just so much coming out where the last year, I think was a lot of a year of, you know, we have horse races between whether OpenAI and Thropic or Google is ahead. And I think that's very, you know, that's good. This sells newspaper, you know, clicks. I think I'm newspaper person dating myself, but the next year is about deep integration into stuff. So, you know, I will be honest with you. even just I play heavily with all models, including open source models, and I spin them up and train them and do all that personally. One of the things that has completely changed my workflow, we are a Google workspaces company and there's products at Microsoft. It doesn't matter, but we use Google and they just cook to Gemini super deep into the amount of context it can have. And I was able to do that before, but just the reduction of friction, the value that we get in our entire business by having very low friction between a model as good as what is kind of becomes the standard that we're all used to using now, connected to the context that I work hard at, even just reducing cut and paste and upload and even more complicated techniques that we use. The single biggest catalyst last year in AI advancement for me was the integration, the reduction of friction between what it needs to understand and how you connect it to other tasks. And I have been blown away. really did, their model got good enough and it got deeply integrated in Q4 of last year. And even just since then in the last month when we're recording, it's amazing how many things have, again, sped up, reduced in time, that compounding gap is growing larger. But it's an amazing thing to think about what that comes back to is the biggest advancement, like let the foundational models get better and well, that's great. They're really good and they'll keep getting better. That's actually not that important. Those will commoditize out. It is truly how this is set up and integrated into systems that will make all the impact. And so that's a really a big thing to think about is understanding the, your business is running and where those workflows are. The integration and seamlessness of that integration is the real productivity bump that you see. speaker-0 (40:21.718) Yeah, and it's to come. think it's going to be such a fascinating field to keep an eye on. And I'm so glad, Josh, that you've had the opportunity here to share some of this thinking with our audience, because I do think everybody should be paying attention. A lot of people are paying attention, but you've given us some great ways to, as both individuals and for our organizations, start to think about the power of AI and getting past agreeing that it would be a good idea if we did something with it and into alignment around taking action. and putting some things into play. I just wanna loop back to our earlier conversation about founders. I think you really offered an important recommendation or idea for people that you have to be able to hold sometimes diametrically opposing opinions or thoughts or suggestions in play and be able to take what, put them in context for you and how you move forward. I think that's true with individuals and organizations with AI as well. You're going to get a lot of different feedback from people about what works, what doesn't work, what you should do, what you shouldn't do. But the truth is this is a fast moving train. And so you've got to find a way to get in the game, find those low stakes opportunities as an individual to really grow your own understanding so that as you're making big decisions, those high stakes decisions organizationally, it's coming from a really grounded place. I really like your four square model around. the unknown, unknown, the unknown, known and so forth. I think that's going to be helpful for people from a decision-making perspective and the CARD acronym. Let's keep that in mind around clarity, ambition, relationships and distribution. As you're thinking about how do I put this technology into play for my organization? All of these steps will help you to get away from just passive agreement that AI sounds like a good idea to how do we actually do something productive with it in our organizations? I want to, as we come to the end here, Josh, just give you the opportunity. If there were just two or three things that you would really encourage people to do as they try to break out of agreement and into alignment around anything, whether it's founding a company or bringing AI to bear or anything else that they might be looking at, what would you suggest as two or three things that you would recommend that people just take on for themselves to try and put some energy behind getting aligned and getting things done? speaker-1 (42:42.766) Yeah, that's a great question. I think I'll start with a couple of things to jump on that the way you've summarized stuff. so eloquent and so clear. really appreciate that. But one of the things that one of my favorite quotes from, uh, or just statements that I always think about, um, came up and building companies in Silicon Valley and, uh, Neval Ravacant is a founder of angels. That's originally he's done a bunch of great things, but he's so clear and concise. always talks about this thing where he's, uh, you solve via iteration. and get paid via repetition. I love that statement in business building. And it's important because there's a distinction between them where iteration is about doing and then failing. Like it's not iteration. You have to have something fail or else you've solved it. You analyze, adjust, and then you try it again. That you requires that. So it's not the famous quote of, you know, it's like doing these over and again is the definition of insanity. This is doing something, analyzing and doing something different. And that is so applicable to. any stage here, but you've advocated some really good ways of opening and failure into iteration to learn how to get better on AI adoption is experiment. And what is experiments? Some fails, some work. You take those, you analyze how to improve, you adjust and you try again. So I think I love that quote because I feel like it encompasses good decisioning. It comes as business building. It encompasses learning. Like all these things are right in there, but then you're looking to the goal to how do you find something that works well? to build repetition into it. And I like that outcome-based structure as well. And you've found a shift to outcome, but a big part of the barrier why people, especially with a low barrier that is AI tooling, as you've so eloquently put, is really just trying it and being comfortable with failure. It's okay not to know and just sort of think with intention. And then I always like, this is something that's come at top of mind as reflecting on now. super related, but it's helpful in the course to iteration is there kind of three ways that I've often think about when I'm putting energy into something, the input, output or outcome consideration. And so if you're looking to say, I want to bring AI into my company, you have to do all the things measuring back from outcomes. Like what are you measuring that matters? There's a lot to that. And you actually have to have a lot of prior knowledge to understand how to impact outcomes. Input is like, I'm going to play with AI for a certain period of time. That's just effort towards it. speaker-1 (45:08.296) and then there's the output layer. The middle is like, how do I, I'm going to try to do something with it, but not actually sure what impact it has. And I think you should start an input, just do, and not worry about too much of the other categories. As you become proficient, you should define, I'm going to go try to do this report and try it see if I can speed it up. That's an output. go move, graduate to that second stage. Then you start to get your grounding under you. And you should start thinking about outcomes where it's like, I think if we can bring this in this way, it can help save this is working back from that thesis. What does that take? And what do need to understand there? And so don't start on outcomes or feel like you need to have all the solution. It's saying, here's the like, first off, let me just use it a little bit for a period of ascribed about time box period of time. I think it simplifies a bit of people's comfort getting into it. And that is not just for AI adoption. That's for anything. use that input. Like, barrier to entry should be input. Am I doing a thing? And you graduate to output. What am I looking to get out of this? Then outcomes, what do I want to change with that outcome? output, excuse me. speaker-0 (46:11.022) And you know, Josh, I'm going to take that all the way back to our very first conversation about your leadership journey. And I just want to say to all of our listeners again, who are starting out in your careers, you're in the input phase. Like, don't worry about, you what's the outcome going to be in 30 years? Like, take everything that you can right now out of just getting going and enjoying this time to, to learn about what it is to be in the workplace and be in a career and be building something together with other people that's really phenomenal. So I think that framework does apply to so many different times of life and also things that we're trying to accomplish, whether they're personal or professional. speaker-1 (46:50.346) Absolutely. just last piece, I'm really quick, I'm telling it to the company. a business is just that building it. It is the process of it not being in the state you want it to be, or even building your career, or building a job, any of the work you're doing in a job. So I love the failure gap because our careers as success should be moving out of failure, covering the failure gap. that is, so building a business, the business is in a perfect state. building a career, it's in a perfect state. is no such thing as a perfect business. There's no proctoring a perfect career. So it's like the process of it, figuring out how to move through the failure gap is everything. So it's like embrace that too. I love the framing that you've been exploring here. speaker-0 (47:35.022) Yeah, it's very emergent, right? Like let's let things emerge. Exactly right. about getting where you're trying to go. Well, we have one last question that we do ask everyone, Josh, and it is just sort of a dream along with me question. If you could get a whole group of people, whether it's your family or your business or your clients or the world or a community that you're a part of, you've mentioned a couple of groups that you're a part of. If you were able to get them all aligned to tackle something together, what would it be? speaker-1 (48:05.806) Man, that's such an important question. it's a good one. I'll be honest. listed this podcast and I thought about a lot of others. heard this question and that's like, I think I came to the conclusive answer, but I would say, I mean, I think the biggest thing that I would say is I spent a lot of time thinking about the business piece of it, but, I would suggest the thing that I admire the most and love and look at myself the best is It is rooted in communication and it's kind of been elevated a bit, but like storytelling is that thing. If everyone were really good at that, which is kind of a question with itself, it's a self, it's like how you get some tale a bit here, but I would say that what I admire about people and even just like my, my family that I'm very tight with when they, we get together, which is that it's bounded and love through storytelling in a good marketing campaign is a narrative told well. alignment on a good, you know, think AI adoption organizations is heavily putted on people being aligned with change management and adaptation and stories. You know, the idea of what you're showcasing the pain point and doing that. kind of is that holistic alignment of what the ability to do that across all cases is one of the things I find us at our best. We're able to align people together. So I'd say that, the thing I always strive and try to work better is telling better stories and getting people more excited and It gets you the spark in people's eye on any topic. It's like if you're, never learned a topic of which I've talked to somebody that if they're excited about it, you get them talking about that, they will turn into the best storytellers of something they have passion on no matter what the topic area. I'd say aligning groups of people into that area of that capacity. I think it's kind of the infinite unlock. speaker-0 (49:49.548) I love that and I think that so there's a call to action there for everybody to level up their storytelling and hey guess what? That's a super low stakes way to start experimenting with AI. speaker-1 (50:00.055) It is. I've literally brought put my trick training program. my partner and I do that all the time. It's like, how do we turn this into a better story? Let's practice. It's a excellent tool for that like it is with many. speaker-0 (50:10.962) Yeah. So if you're thinking to yourself, well, I don't have a way to experiment. We've just given it to you. Figure out how to tell your stories really effectively and in a compelling way and use some of these technology partners to help you to do that. Josh, this has been such a great conversation. Thank you so much for sharing yourself more about what you and your company are doing and really hoping, hopefully inspiring people to move through the failure gap when it comes to things like AI tools, but also when it comes to tackling big ambitions that we have in life. I really appreciate some of the tools and resources that you've referenced. As I said, I'll put a lot of those in the show notes too, so that people can find them if they're interested. And I just want to say a big thank you for being here with us today. speaker-1 (50:53.23) Thanks so much for having me. This was such a fun conversation. I love the topic area. Thanks, Julie. speaker-0 (50:58.126) Yeah, it's a great conversation. And I would say to all of our listeners, if you enjoyed this episode, please remember to like it, share it, comment on it, reach out if you want more information, and we will see you next time on the Failure Gap.