← Using AI at Work: AI in the Workplace & Generative AI for Business Leaders
Using AI at Work: AI in the Workplace & Generative AI for Business Leaders
93: Using Generative AI to Develop a Winning Strategy for Business Leaders with Justin Trombold
00:47:35
Loading summary
Transcript74 lines
- [00:00]
A
Often you get a different type of burnout from doing generative AI over and over and over again on the same thing.
- [00:07]
B
What are the commonalities that you're hearing? Where they're getting it right, but also where they're getting it wrong.
- [00:12]
A
There's a train of thought that treats it more like an IT project when it's actually more of a transformation project.
- [00:20]
B
So how do I navigate this? Hey IT guys, we need you, but we don't need you need you.
- [00:24]
A
The short answer is centralizing the decision making from a tool perspective, but decentralizing the decision making and deployment from a use case perspective.
- [00:33]
B
I mean, you give them what they want or do you give them what they need?
- [00:37]
A
They have to get their hands dirty. They have to be fully immersed in using AI in their own work.
- [00:42]
B
Justin Trombold is the founder and president of Antisen Advisors, a gen AI strategy leader who helps companies answer the big question, what are we going to do about AI? Moving teams beyond pilots into measurable results through readiness, diagnostics, practical workflows and culture. First transformation. Welcome to Using AI at Work. I'm your host Chris Staigle. Each week we'll be learning how today's business owners, entrepreneurs and ambitious professionals are getting more done with smart use of tomorrow's tech. Let's get started. Right now, every business leader is asking the same question. What are we going to do about AI? If this is you, chiefaiofficer.com has the answer. We give you a simple path forward where we provide executive and team training so your people know exactly how to safely use generative AI in their day to day. We also manage the deployment and implementation to make sure tools actually get adopted and deliver results. And we'll also guide company wide transformation so AI becomes part of your operating system, not just another shiny object. The companies that act now will increase productivity, cut costs and grow faster than their competitors. Those that wait will get left behind. So if you want to make AI work in your business, visit chiefaiofficer.com and see how we're helping companies of all sizes finally get results from AI. Hi everybody. Welcome to another exciting episode of Using AI at Work. My name is Chris Daigle and I'm the host of the show. And today our guest is Justin Trombolt. He's the president and founder of Antisen Advisors and he is talking to a lot of businesses that are asking the question, what are we going to do about AI? And I'm always excited to bring other people who are out there, like not just coming up with ideas, but actually dealing with clients in the wild when it comes to hey, we gotta do something, help us. We've got these ideas or we've got these misconceptions or whatever. And there's a couple of threads that I'm gonna pull on for sure, Justin, but before we get started with that, maybe just give a little introduction about what your career has looked like that so that you've arrived on this podcast today.
- [02:43]
A
Yeah, and I love the way you said that of being out there in the wild. Cause I think certainly like that more than it's even felt like before with consulting and the work that I do. But I'll keep a very long boring short or long story and boring story short and perhaps still boring. But just so people know who I am. But I was an academic teacher and researcher for 10 or 15 years in biological sciences. And you're doing a lot of various different work. Your listeners might be thinking, well, why is this guy now talking on an AI at work podcast? I about a decade ago I switched and started a new career path working in consulting. And so your typical management consulting, you know, big four strategy firms that those types of cultures had a chance to work at a number of those different firms and over the last three years transitioned out of that lifestyle into more independent consulting. Some of that had to do with just my proclivity for liking to go my own way a little bit. But some of it, I have a three and a four year old at home and the big firm consulting lifestyle, while independent consulting is busy, it's a different kind of busy. So I have a little bit more mental bandwidth for those guys. And I think, in short, there are a lot of things that I can do. But the relevant part for your audience is I work with a lot of organizations, whether it's your larger Fortune 500 type organizations, all the way down to, let's say 5, 10, 10 person family shops that do something and even solo solopreneurs, individual entrepreneurs helping to understand not just what generative AI is, but a way of thinking about generative AI and AI in such a way that they can consume it in the context of their business and deploy it in the context of their business. So that, that takes, yeah, that monster takes on a lot of different shapes and sizes. But there are some central tenets that string together that I find are true for the biggest companies you hear of down to again, a single individual that's trying to build something. So it'll be an interesting conversation and a lot of different experience in different industries. Both in and out of AI. So excited to talk to you about that, you know.
- [05:04]
B
Well, I think that's a great place to start. Like, what are the commonalities that you're hearing? Where they're getting it right, but also where they're getting it wrong. Like your clients or these prospects that are interested in getting help with AI?
- [05:17]
A
Yeah, I think the easy one, you know, there are obviously companies that are getting certain things right, but we'd say, like, well, what's the common thread that people are getting right? And it's usually the fact that they have. This is something they have to pay attention to. Right. They have a degree of focus on now that like most things, putting a little bit of focus into something is very different than putting a lot of focus into something. And when you put a lot of focus into something, you're always at risk of over investing or misallocating resources in certain ways. And of course in this case, getting confidently wrong answers or perhaps getting lazy in the type of thinking you're doing and so forth. But so when organizations get right, I think is the understanding that there has to be adoption. Now, what typically gets wrong? And we can pull on a bunch of different threads, but I think one area that's notably, I'll say two that are notably wrong is one, there's a train of thought that treats it more like an IT project when it's actually more of a transformation project. So that's one part and then the other part where there's a lot of mistakes and it's, it's a little bit of a different flavor of the first one. But it's the concept that there is this technology solution, but it isn't a technology solution that can solve every problem. And what most of the time organizations get wrong is they assume if we plug in solution X into workflow Y, we're going to get outcome Y better and faster than we did before. But what usually happens is it improves some aspect of process Y. And so maybe that's accelerated and that's improved, but the process itself just gets stopped basically at that point. And we can dig into that in more detail. But just think of an idea where a typical bottleneck in a process, oftentimes what AI will do is it'll just shift the bottleneck to a different part in some sort of workflow. So that's a little bit more of a detailed version of the first one. But yeah, the same. You know, it's, it's change is one and two, it doesn't solve a whole process or a whole problem. It tweaks certain aspects of a process or problem.
- [07:38]
B
Okay, I like that a lot and I am going to dig in on that. But I want to start with this, this IT conversation because for. Okay, I'm an executive. I've got maybe a free account to ChatGPT. I'm kind of using it. I know we've got some people scattered throughout the organization who, I hear chatter that they're power users, but I have no idea what they're doing. I'm too busy to sit down and watch a bunch of YouTube videos. I've heard some, you know, some presentations at our industry events over the past year. And, and I'm intrigued. But like, I've never really got gotten, quote, unquote, gotten what AI is. So to me, as that executive, it seems like, hey, bring the IT guys in here, let's talk about this. It seems obvious, right? This is a technology, it's artificial intelligence, machine learning, data science. But generative is not that. So how do you. Because here's what I don't want to do as the executive. Obviously I want the right information, but I don't want to alienate or I don't want to create this, this contrast between, oh well, it's not the IT guys, they're going to feel left out. And I know how protective IT is of, you know, their role and their domain. So how do I navigate this? Hey, IT guys, we need you, but we don't need you. Need you.
- [08:56]
A
Yeah. You know, going back to the Dilbert cartoons, you know, it is always a point of some, some jokes and so forth. And there's always a nugget of truth in that. I, I think it's. As a leader you can set the precedent we see like, with like anything else. If you set the precedent early in your organization and perhaps as a leader, you know, some listening has already done this. But what is if you make it an IT first priority and you bring in the IT leads and put them in charge of the, let's just say the generative AI agenda that, that starts with what you typically think of, of selecting tools. And that very much could be in the domain of the IT team, right? Of which tools can you can and can't you use? And security parameters, so picking the tools. But then what we see is that the, the, the further the IT team starts creeping down into use cases and deciding where investments should and shouldn't happen, that's where you start to get a fundamental conflict. And it's a conflict for a couple reasons. Because There is an interpersonal aspect of it. You're going to start seeing solutions that are deployed that don't necessarily address business problems. And you're going to see what you typically see with when anyone, any business gets a new technology solution, they're like, like, this doesn't have anything to do with anything I do. Like, I don't care. So you get that natural friction that will start emerging. And what we see in that case is the people that you really need to have engaged. And it isn't. We were talking about this a bit before the conversation. It's a mindset of wanting to empower and engage your people and make them feel like they have and for them to truly have ownership over how it's being used, where it's being used. And so what we typically see that works very well is you do have an IT thread. Right. But that's more about the toys that are in the sandbox, right?
- [11:01]
B
Yep.
- [11:02]
A
But in terms of, let's say, picking which toys are getting played with, how big that sandbox is, what games being played in Sandbox, you know, whatever, you know, everything that you're doing. So that's an analogy for deploying the tools. What we see is having an individual in charge at the business unit level, you know, somebody that's in the business. Yeah. So this could be a business unit lead, a functional lead. In a smaller organization, it could be. Well, this is the person that typically manages HR or typically manages product for our company. Whatever the roles are and however it's organized, you make them the lead and what they do. And this can show up in a lot of different ways. But they're the person that makes sure there's connectivity between business problems and the way the tools are deployed. And on occasion, there'll be an upstream conversation. Hey, we need a different tool in that case, of course, there has to be some conversation, but there shouldn't be a conversation in terms of what the use cases are unless there really is a true feasibility, technical feasibility bottleneck, which is almost never the case in a lot of these tools. Yeah. And so the short answer is centralizing the decision making from a tool perspective, but decentralizing the decision making and deployment from a use case perspective.
- [12:30]
B
Yeah, that makes sense. Okay, so it's involved when it comes to. Now, here's here would be my concern if I was, you know, a listener to this and exploring this. Where, you know, most companies are in Microsoft. Oh, well, we've got Copilot baked in and Copilot has all of these things. They've got a partnership with OpenAI, they've got access to Claude, now we've got agents. However, our experience in working with clients that are in the Microsoft environment is that that tool as of today is falling short when compared to the other options that are out there. And as an executive, I don't know that. Right. Because it just all is the same to me. Let me ask you, has that been your experience?
- [13:14]
A
It has a little bit. To be honest, I do find it's often a little bit more of a theoretical question than one that comes up too often. I'll say that in larger organizations where they have, let's say, preferred vendors and they have a stack in place provided that there's access to some form of LLM and you know, there's, there's some trade offs with different ones but let's just, you know, put them in that. You, Gemini, Claude, chatgpt, grok, you know, the four or five that are the core, as long as you have access to those and you have enterprise licenses and perhaps there's, there's a reasonable amount of restriction on what you can and can't do in using the model. Right. So obviously if the restrictions are too high, that can create another set of problems. As long as you can clear that hurdle as a person in the business, you just, you might have to be okay with, look, I'm not going to win this battle. You know, we're always, we're typically, you know, we go with this type of vendor, we have relationships that led, yeah, led to us having anthropic instead of, let's say, let's say Grok or Instead of your ChatGPT or Gemini, it's more of just accepting that as a reality and then just getting used to using that system. I think what I found in my own work and working with clients is once you spend a half day with any of these models, you almost forget the model that you're working in and so you get past that. But at a smaller organization you have a different type of problem which is oftentimes there's a hesitancy just to get over that first hurdle. Particularly if it's, let's say a family shop and the leader is, let's say a more senior person, hasn't really been around much and maybe they're uncomfortable with even deploying an enterprise license of ChatGPT for enterprise use. That's a different argument. So these are all more political aspects, but if you can clear that LLM bar, you can do most things you would need to do in the way that at least we talk about now, you might be restricted in certain applications you could purchase that may or may not talk to those models. But in terms of deploying those LLMs and exploring use cases and testing and scaling, you could almost do anything with any of them that's technically feasible.
- [15:43]
B
At this point, I'm satisfied with that answer. And I'll tell you what I just realized is that I'm, because of my, like, I guess my, my depth in this subject, I'm jumping the gun. Most of these people, like, this is their first experience with any type of LLM. So the reality is, you're right. It doesn't matter what they start with. Once they exhaust the capabilities of that model, then maybe we talk about, well, this, let's add some more, let's get, let's get Chad GPT enterprise licenses or whatever. But I think, I think that's really good advice is don't get caught up in, oh, but this one's better at this. And this one's just, you know, ride the horse you got. Right. I think that's great advice, actually. Okay, so who needs to be in the room when we're ready to have that conversation? I like the idea of the business unit or the domain kind of heads because they're going to talk about the problem from the bottom up approach. We know the pains that are happening with the people on a daily basis, but we may not be thinking about the strategic alignment of the use cases that we think are important. Right. Yeah. So how do I, when we're ready to say, okay guys, we got to do something, let's figure this out. What is that? How are you structuring that conversation with clients?
- [16:58]
A
Yeah, so it's, it's one of those things that's very easy to say and always kind of hard to do. But, but the two, the two buckets. And it's always, it is always helpful to clarify what it is that's being discussed. Right. Because if you're having a discussion about how to use an LLM and it isn't clear if you're talking about, let's say, realizing an enterprise strategy and some broader vision versus solving some more precise problem, it's going to be an impossible conversation. And so the way that we, we sit around this is split it into two areas. The first area is, okay, we have our enterprise strategy, we have our business unit strategy, we have some strategy. And it could be at a more local level within the company, but you have that strategy. Then you start asking yourself the questions of within a given strategy that we have, where are those opportunities to leverage generative AI to help realize and facilitate that strategy? Now, there's a lot that goes beyond that, but just conceptually it's good to think about it as separate bucket. Now as you get further along, that can be reimagined where you go back and say, should we change our enterprise strategy because of those tools? But it's premature to have that conversation before you've even deployed anything. You have to have your hands in it and understand what it can and can't do before you do that. But the second bucket is okay. Person X, Business unit X, Team X. How can you deploy generative AI tools? And the way that we recommend is always starting with more of an open forum LLM approach. So going in and deploying those tools in your work to first maybe solve a specific problem, accelerate something. And as part of that, a key thing that I always like to talk about, and we alluded to this a little bit earlier, is that start with something simple. Start with one thing. Maybe it's one process. It's one thing we do. Ideally it's important. It doesn't necessarily have to be, but ideally it's important. Map out that workflow. Find the place where you think. You don't even have to know for sure if it's a generative AI solution. At that point it's just, hey, this is something that's very manual. Yeah. It's very text heavy. Those series of questions to identify a play. Or you could, you could even put in an LLM and say, here's my process. What's a candidate? Right. You can use. So, so you find that point. Yeah. But you have one step before you start experimenting and that is okay if this was faster or if we got this different output or result from this step in the process.
- [19:49]
B
Yep.
- [19:50]
A
What has to be true downstream or upstream for that to be realized? Right. And so like, yeah, the example, you know, we were working with a client and I think it generally just resonates even if someone isn't even, even if someone isn't in sales because it conceptually makes sense. We were working with a client to deploy a, a lead qualifying and lead, you know, qualifying, scoring. So it was about like, is it good? Can you warm it up a bit? Like, how do you create custom content? And what that was doing in that case, that was asset management. Okay. In that case. And so these leads, this, this flow is coming through and. But this holds. I think it's a pretty simple. It holds through to you know, most anything. And yeah, the leads did accelerate, but what wasn't different is the salespeople either didn't know how to use what they were consuming, or perhaps more importantly, there wasn't a change in the incentive structure in order to incentivize the salespeople to sell more.
- [20:53]
B
Yep.
- [20:54]
A
And so, and in that business with that company, they were more incentivized to expand existing accounts than to sell the new ones as well. And so they just got this, this blog jam of qualified leads. And so ROI from that solution. Absolutely not.
- [21:13]
B
Right.
- [21:13]
A
Because it's not going to do anything. So that step there that's in between is look at it from a non generative AI lens, from a first principles lens and say, okay, the bottleneck was maybe in finding good leads before, but now it's in actually selling to the leads. What has to change to now address that new bottleneck. So you do that, then you have, it could be an out of the box solution, but it's more likely that you're going in, particularly with smaller companies and you're like, okay, well we have this process of qualifying leads. Let's start experimenting with putting together really good prompts. So we work with clients a lot. What does a good prompt look like? How do you then actually use the LLM to make the prompt better and better? Because you can do that as well. And then how do you get better at directing it to good information sources? You know, that's a very important thing because I'd advise your audience just Google the typical information sources for any of the major LLMs and that'll make the case for me in terms of where that comes from. And then they start experimenting and testing. But what's key is there has to be a very clear set of parameters that are set up on the front end. It's not just about what models you can use, it's also about, well, if this starts working well, what does this look like? What does it mean? Yeah, well, what is it? What even, what does it mean for the person that develops the solution? If it works, how do we know if it works? If it works, is there a light at the end of the tunnel in terms of scaling and investing in, let's say a custom agentix solution that does it really well so you're not reliant on that person. And so it's two categories. So you have this idea of deploying it for the strategy and you have this idea of what the individuals are using. Now you can have connective tissue there. Right. So back to that sandbox. A good generative AI strategy is linked to an enterprise strategy or a business unit strategy.
- [23:23]
B
Sure.
- [23:25]
A
And within that, that's what sets the walls of the sandbox that you're playing in.
- [23:31]
B
Okay.
- [23:32]
A
And so then what often is true is that the individuals that either in the center of excellence for AI, if it's a large enough company, or the leaders for any size organization, they often won't have a good vision into how that has to come to life or like what needs to be true for that to work. We're often prior, you see clients on the receiving end of that, prior to generative AI on the receiving end, where it's like they want us to do what? Like we can't do. Like that's not possible. Like we, they don't understand how we work. Now it's a little bit different of you want us to do that. You don't understand what it is we can do. Yeah, yeah, yeah. You know, but, but the people sitting on the ground do. As long as there's clear communication, right? Yeah.
- [24:17]
B
So this gets tricky then, because, you know, like, if you think about it, well, we've got this one pain point in the, in the company, we don't like doing this activity or we've got a lot of people that spend manual time doing this. That's probably what a lot of people have heard. Find the, find the manual. You know, where you're spending a lot of time manually doing stuff. But just because we solve that, if we haven't considered upstream or downstream from that fix, then like, okay, now we've just opened up the faucet, but it's flooding this other area of the business that wasn't, that was, that was built and designed to accommodate what we were able to do as far as throughput output constraint. Interesting.
- [24:58]
A
So, and I'd say one put one thing on and that, that challenge that you talked about of, of the downstream becomes harder and harder if you start crossing borders, whether it's between business units or between functional parts of the business, that you'd imagine that if it's just you working like that bottleneck, you're very aware of that bottleneck and everything and you're not stepping on anyone else's toes and you're not reliant on somebody else. And so one of the key tenets in the research that we put together and this has been bearing out in client work is that the ability for organizations to collaborate cross functionally and not work in silos is a big driver of it isn't Necessarily like a single use case, if it's a constrained use case, to like a small group. But anything that spans business units, spans functional units. That if you're a siloed company and you have an ownership culture, not in a good way, but like, this is mine, not yours kind of ownership culture, you're going to see a lot of problems that start to come up when you do that. And it'll. It's not going to break the silo down. It'll actually fortify the silo because people will start.
- [26:14]
B
Interesting.
- [26:14]
A
We'll start feeling more pain because of each other. So it's. In some. What we've seen in some cases is there's an assumption that if the other team, let's say, starts to see. If the salesperson starts to see more leads come, they're going to then want to see. They're going to want a solution. Right. But what happens in some cases is, well, you're giving me all these leads, but I want to talk to my existing relationships and you're putting pressure on me to talk to new leads. Like, now I'm upset. Like, I'm not just not more efficient. I'm actively upset by this thing that changed and is now messing with the way that I sell work.
- [26:54]
B
Interesting, man. We're covering process change management. We're covering all the stuff here. But these are all important because I think that maybe the listener. You're the listener and you're checking this out. You're like, hey, we're ready to do some stuff. And you kind of think it's going to be this fun process of, ooh, look, A.I. magic. I was able to, you know, write the email or the very basic things that people get started with. And you, you don't know what you don't know. So as a result, you kind of get started with those things, but you expose the company to risk because there's no use policy in place. There's no. Hasn't been really any training. The training came from somebody that watched a TikTok video and is now doing the thing in their role and you start doing that. And then there's this. Well, the. They feel like my. Their toes are being stepped on because all of a sudden you're forcing me to do something that wasn't part of my job description or what? Like, this is really getting tricky here. So how do we make sure that when we're getting started that we are thinking we're playing some chess moves ahead so that we're not just getting caught up in the Magic and the, the. Oh, that's cool. But we're actually doing this considering the downstream impacts.
- [28:07]
A
Yeah. And I'll do a quick sidebar. Yeah. One of the things we work with clients on at the beginning of whether we're talking with them in a training session or we're speaking more generally about what has to be true in some of these operating model designs, there's a mindset that goes into it and it's important that the leaders and individuals throughout the business have an exploratory, have a curious, have an excited mindset about doing this. Because if they don't, you, you, you, you just, you aren't going to get the same level of buy in and the same because it is fun when you have the freedom to do it. So just setting the idea that we're going to give you some space to explore and we worked with a client that they had this culture in place a bit already with prior technologies. They just applied it to generative AI where they're mandated, it's a loose mandate, but they at least give them the Space to deploy 20% of their time to integrate generative AI solutions in their existing projects. And so they don't have to do it, but it's there. So again, that's about mindset. But to your question of, yeah, a bit more specifically about, well, what do you do? There really are two paths that we see are useful to take. If you're an organization or in an organization that likes to understand a little bit more, more of a diagnostic activity and think about, well, where are we going to see some of the challenges? What types of use cases are we more likely to be able to deploy based on the way we work and everything like that. Start with a diagnostic activity to assess your generative AI readiness. So we do a simple one that goes through five pillars. The first one is alignment with your strategic vision. So generative AI vision aligned with strategic vision. Or you could say just aligned with a business problem. If you want to make it simple, then it goes down and gets into end user proficiency. Then looking at cross functional collaboration like we talked about, there's scalability and adaptability which is are we willing to make some decisions and change some things if we see evidence and then governance and regulation or compliance. So some companies like to start there and it's a nice place for anyone to start, but it's an academic exercise in a lot of ways because it's still a. Then what? Well, what we recommend and what we work with clients on is even within a business unit. Take it from this macro thing where everybody's involved and let's say the VP of that business unit is in charge of that or whomever and shrink the problem down into a small group or even just an individual, maybe two individuals, and go through a process where again we talked about before, pick one workflow, pick one part of that workflow, define one KPI that isn't roi. Do some exploration and then see if that KPI is sufficient to understand if it's working. But it's maybe created another problem. So like in that example before, if you just track number of leads generated, you're going to see an effect, but obviously there's a problem downstream if they're not, if they're not converting. So pick one, maybe two KPIs that aren't ROI.
- [31:41]
B
Like it?
- [31:42]
A
Yeah, again, yeah, like, and so in the case of the sales, it would be if you saw, if you were tracking the number of qualified leads plus the number of new accounts being sold and maybe even you added in like the time spent with existing accounts, you would probably see this thing where the time with existing accounts might be going down, the lead flow is going up and there's no change in the sales volume. And so you don't want to overcomplicate it, but some sort of simple set of KPIs, this is where that shrinking comes in. Have a relatively simple discussion about what has to be true within the way you work. So let's just say that that was a very small company and the sales team is one or two people. Right? That's a very simple conversation of just saying, well, okay, we need to talk to leadership and we need to change the way we're incentivized to be able to sell new business. And so let's have that discuss. So make the operating model changes that need to be done and can be done. Right. Oftentimes at this stage what you want to be, you want to be had the teams be flexible on is be willing to say, look, we just can't do this right now. We're not, they're not going to change the incentive structure. Like we just need to go down the path. But yeah, get to that point and then create it. Sounds old fashioned. You know, it's an old business business word that probably everybody, everybody hates. But put together a nice simple charter that just says, here's the experiment, here are the KPIs, here's the owner, here's the decision maker, this is what success looks like. And then run, you know, a 7, 10 20, 30 day experiment and then have there also be a clear path of, well, if success is met. What does that mean? What does that, what does that look like? Is there a commitment to invest in this solution? So I want to tie this back to the question that you had.
- [33:41]
B
I like that.
- [33:42]
A
Yeah, well, what's good about it is makes it to where you can have all these little micro investments. Yep. And micro decisions. And they may or may not be aligned with this broader corporate strategy, but they're certainly aligned with the general principle of you're trying to make things more efficient, maybe improve ROI or functionality of a given unit. So you can have a lot of these micro pockets. But the key is start small, keep it simple, find solutions that work. What happens then is the individuals start to see it improve their own work, the more terminal leaders start to see it work, and that energy starts to propagate upward and outward in an organization.
- [34:28]
B
I like this idea a lot about a document of, okay, we're going to do this. What are we looking for? What exactly are we doing? So there's not scope, creep and ooh, let's do that. And then how do we know that we've won? How do we know that this has been a successful effort? I like that idea a lot. As a matter of fact, we're going to start introducing that into all the work we do because it just seems.
- [34:50]
A
I challenge your listeners just to think. Yeah, well, think. Think for a moment. And maybe for your listeners, you ask yourself, are you someone that feels like AI is improving the work that you're doing or making yourself faster? And okay, if you say yes, say, well, what are you basing that on?
- [35:06]
B
Great. Yep.
- [35:07]
A
Now it might be someone's very, you know, a very conscientious person, and they have it, they have it all tracked and so forth. But it's at least a coin flip that when you ask yourself that question, you're like, what? I think it, I think it makes things better.
- [35:24]
B
Gut.
- [35:24]
A
Yeah, I'm pretty sure it does. You know, but, but does it? And so, you know, having that and. But, but the other element of it is it makes it to another pitfall that we see that can. Can really derail enthusiasm for, for generative AI is you have people doing experiments. They just die on the vine.
- [35:42]
B
Yep.
- [35:42]
A
And that person can keep doing it and perhaps they're okay with that. But often you get a different type of burnout from doing generative AI over and over and over again on the same thing. You know, you can start to become very frustrated when you know that there are solutions that you could, you know, bring in a vendor and they could build, let's say, a custom agentix solution that puts all that together and really streamlines the workflow. So keeping that vision and excitement. And the last part, that it's a bit of a tricky conversation because it's a different way of thinking about work. How do you incentivize this type of behavior beyond just telling your people, well, you're preparing yourself to be a leader of the future. It's like, okay, well, okay, that might be enough, but it might not for your people to encourage that type of enthusiasm and exploration.
- [36:38]
B
And this is opening up a lot of questions because we do a lot of like working with clients. And just when you think you've like, oh yeah, we got it, we got it figured out, you bring up a couple of things here that have certainly exposed some obvious gaps in what we have been doing that are simple fixes but will have big leverage. In particular, I like this. Like, I realize one of the things that we're not doing probably to the degree that we should, is starting out the relationship with that, that strategy evaluation. So what does that look like for you guys? Do you say let like bring out everything and let's take a look at it or what does that strategy conversation look like for you? That, that initial combo?
- [37:24]
A
Yeah. And you know, I think anyone that's been in any sort of client services work knows that you can have a central theme of what you're trying to do, but it takes on different shapes and sizes with different clients. Right. So the general theme of it is a way to look internally and explore each of those five areas. And so what we have set up is a formal diagnostic survey that ideally you would deploy it across the organization. Obviously you probably wouldn't get everybody, but the more people you get, the better because you get a better view. But it's basically exploring what would be maybe red lights, yellow lights and green lights. And what we see with organizations, and I have a white paper that maps this out a bit, is that there are different generative AI readiness Personas or phenotypes or whatever your audience thinks about that stuff based on that Persona, there are different types of use cases that are going to be more or less accessible now. Yes, the utility in that is, and this is more so with the how do we realize our enterprise strategy more effective? Because you can say, okay, well where can we win today? Right? What are the things that we won't have as many problems with because of how we how we work you. Yeah. And then you set up a plan for where what you have to do or what you have to change to win tomorrow. Right. So those are, that's getting into that transformation element, that operating model element, which is what no leader in any business wants to hear, is that the problem's in the operating model because that's the hardest thing to change because it's transformation. But then, you know, once, once you figured that out, this other process, that second process, we talked about shrinking it down, you can overcome a lot of those barriers just by making an experiment smaller. Right. So that type of exploration. Yeah. It's not going to have the same type of impact. But if you think about what the alternative is. Yeah. Do you want to invest a million, five million, $10 million in an enterprise use case and an application or a tool and your organization just isn't ready for the implication of what that means. So you know that, that survey is great. Oftentimes it ends up being a bit. It can be like a 10 minute discussion with the CEO of a company. It just, it depends on the company. Right. You know, sometimes they're like, look, we're okay, we're okay here. Like, yeah, we need to improve X. And you know, then you run with that. But yeah, it's starting with looking internally, looking in the mirror first and then going from there.
- [40:14]
B
What do you do in situations where, because my personal position is that, yeah, you've got that pain point here and you've got that pain point here. We can come in, we can have an automation or an agent address that. But that doesn't prepare your company for this new landscape that's happening. Right. Like, like that's not AI. That's one small sliver of it to me. It's your people need to be trained, they need to be doing what we call thinking in AI to where it's just the default for them. Right. They're gonna like, if there's a problem in the business, they're like, how can the models help me? Or is there an AI tool? Right. And if you don't do that, you can have all the tools, but your people aren't ready for a competitor who does have AI fluent people who are doing incredible things simply with just a co pilot license.
- [41:00]
A
Right, yeah.
- [41:01]
B
So how do you, how do you, I mean, do you give them what they want or do you give them what they need?
- [41:10]
A
Well, I think I'd say, and we were talking about this a bit before is, you know, training does have to be thought of differently than what we typically do and you know, it, it could be, there could be modules internally and so forth, but to get the ball rolling, but it has to be ingrained in what they're doing. They have to get their hands dirty. Yeah. It has to be, they have to be fully immersed in using AI in their own work.
- [41:36]
B
Yes.
- [41:36]
A
And what the target is, if you think about for your listeners that maybe are more fluent in AI and for those that aren't, we could talk about it another way. But if you feel like you can problem solve in an LLM, then you're kind of where you need to be.
- [41:56]
B
Great.
- [41:56]
A
Right. But if you feel like you're sitting. Yeah. If you feel like you're sitting down and you don't know enough to, let's say think to ask the LLM. Well, where within this workflow I just defined is a good generative AI application, then you're probably not, you're not, you can't critically think in LLMs yet.
- [42:16]
B
I like that.
- [42:18]
A
Yeah. You'd need those boots on the ground experiences. Yeah. And you need to kind of organically export and so this is where I've heard it said in a lot of different contexts, so I can't attribute it to anyone in particular. But how important now is curiosity over pure intellectual horsepower? I would say, I don't know with absolute certainty how much more important curiosity is. But having people that are curious and want to get in there and not just explore, but their mind naturally goes associative thinkers or abstract thinkers, you know, they, they often are individuals that it seems like are adopting these solutions makes sense. At least, at least they're, they're doing it a lot, you know, and they seem to be doing well. But that's, that's supposition. I don't, I don't have any data to back that up.
- [43:10]
B
You know, I like it because the individual of high intellect may have bias. Right. And the curious individual is like, well, what if we did this, what if we did that? And that's where the, the true like aha's come from. Using the models is the curveball stuff, not the standard. You know, write me an email for a subcontractor who has missed their deadline, like helpful. But that's not what we call thinking in AI. Right. It's fantastic.
- [43:37]
A
Well, I'll just share this. I know we probably have to wrap up here in a minute, but I'll just share an anecdote with myself. I mean, I think of a specific moment where I was up late at night trying to Work on something. And it got to the point where the LLM was started, instead of answering my questions, was giving me tips for coping with disappointment because it wasn't working and I was getting noticeably agitated with the LLM. So going through those motions, that's how you start to learn. And if you're, if you're not having those types of experiences, you're not learning, you know, the paths and the roadblocks to using a tool like that. Yep.
- [44:26]
B
No, that's perfect.
- [44:29]
A
Wow.
- [44:29]
B
There's so much more that we could talk about, but this was good. I've pulled a few ideas I'm going to bring back to my team and say, why are we doing it this way? That's pretty good, Justin. So you mentioned that you've got some white papers and things like that. For those who want to dig in a little bit more with, with how you're approaching this stuff in, you know, everything from, like you said, from SMB all the way up to enterprise. Where should they go to get more of this insight from you?
- [44:54]
A
Well, I don't want to sentence anyone to have to read my white paper, but they can read my white paper. It's, it's, it's on our website. I'll perhaps share it here for the show notes.
- [45:03]
B
We'll have it in the show notes.
- [45:04]
A
Yeah, yeah. If you want to get into more of that question of, you know, what are these five pillars of readiness?
- [45:10]
B
Yeah.
- [45:11]
A
How should you think about different use cases that can be deployed? How do you think about that? When? Today, when tomorrow? Roadmap. It's all in the white paper there. But probably the funnest way to engage with us is on our website too. And I'll provide this link. We have like a 5, 6, 7 question miniature version of the diagnostic survey and so you can go in. It doesn't, I believe the full one's like 25 questions or so, but it pulled out the ones that had the highest relationship to readiness and crunched it down. But yeah, just go and explore. Just it is even really about scoring it. It's just seeing the types of questions that are posed and thinking about those for yourself. And then you can submit your information and we're of course happy to have conversations with any of your listeners.
- [45:58]
B
Awesome. Well, man, thank you so much. I know that if you're in the AI space as a professional, it's extremely busy times for all of us right now, so I appreciate you making the, the time to come in and talk to our guests today. We're going to have links to all this stuff in the show notes and I would encourage you if and where
- [46:16]
A
are you based in in the Charlotte, North Carolina.
- [46:20]
B
Okay. So and obviously I mean you can work wherever, but if you're in that that region it would be my preference is always to work with with clients locally just because there's something that the zoom is one thing, but sitting in the the room with them on the whiteboard and the the laptop open is always a good thing. So. Well, awesome. Justin, thank you so much again for the time and for all of our listeners. My final bit of advice every single time is just go use the tools. Go use AI. So with that, we'll wrap it up and we will see everybody next week with another amazing episode of Using AI at Work. Thanks Justin. Thanks everybody.
- [46:55]
A
Sounds great. Thanks Chris.
- [46:57]
B
Thanks for tuning in to Using AI AI at Work. Don't forget to subscribe for more conversations about how to use AI at work and a special thank you to our sponsor, Chief AI Officer for Empowering Businesses with AI Education and Training. Visit their website for a free AI Readiness Assessment and AI Strategy Guide to help you get started using AI at Work. That's www.chiefaiofficer.com. follow us on Twitter at the handle Using AI at Work and visit www.usingaiatwork.com for free resources to help you harness AI in your.