Getting Your Team to Use AI with Rick Lindquist

Rick (00:04.025)
Hello, I’m Rick Lindquist. Welcome to a special episode of the Growth Elevated AI Sherpa podcast, where I will talk with inspirational entrepreneurs and leaders in the tech industry. Some past guests include CEOs and CXOs of great companies like Workfront, CHG Healthcare, Pathology Watch, In Moment, Canopy, the San Francisco 49ers, and many more. This episode is brought to you by Growth Elevated, a community of tech founders, CEOs,

Julian Castelli (00:10.224)
I sure don’t guess.

Julian Castelli (00:15.043)
Yeah, okay.

Julian Castelli (00:26.594)
This episode is brought to you like Rip Hell data. Or did it keep working under your practices? Yeah, expensive podcasts.

Rick (00:33.079)
CXOs, and more who are committed to working together to share best practices and learnings in an effort to help all of us become better leaders. We do that through educational programs like this podcast, as well as our blog and of course our annual ski and tech summit, where we bring tech leaders to beautiful Park City, Utah to enjoy camaraderie and collaboration in the beautiful mountains. So if you enjoy skiing in the outdoors and networking with other tech leaders, check us out at growthelevated.com.

Julian Castelli (00:49.66)
So if you look at the working with other tech leaders.

Check this out and road related described as a password gas. Hey, I’m excited to be talking to Kelly, did about how companies can adopt way that the number of issues. Julia is an accommodate corporate executive, founder, and now an operating who works with multiple staff and

Rick (01:02.635)
And please subscribe to this podcast wherever you listen to podcasts. Today I’m excited to be talking to Julian Costelli, the founder of Growth Elevated, about how companies can adopt AI successfully within their organizations. Julian is an accomplished corporate executive, tech founder, and now operating partner who works with multiple SaaS and tech companies as the chairman of the board and active operating partner. Julian serves as a key liaison between the executive teams.

Julian Castelli (01:26.121)
And have the top requirement. people context piece and

Rick (01:31.019)
And the financial investor and sponsor helping to orchestrate constructive value-added board process and value creation. Julian recently launched an AI adoption workshop for the growth elevated community, and today we’re going to review stage one of that process. Julian, I’m excited to talk to you about this today. can you tell us a little bit more about the context for this and what you’re building?

Julian Castelli (01:54.036)
Hey, thanks, Rick. glad, glad to be here today. as you know, AI is the topic de jour. It’s it’s been for the last couple of years. I suspect it’s gonna be for the forward few years. And everyone’s talking about AI, but there’s been a lot of FOMO. In fact, that’s the whole idea behind this podcast is to to try and get out there and share what’s happening in and practical use cases. And I I was talking to the people that I work with, the executive teams where I where I’m a board chair.

And just other people in the community and everyone was doing something a little different. and almost nobody felt like they were totally on it. I’ve I I rarely talk to someone said, yeah, that AI thing, I’m I I’m I’m absolutely killing it. I’m I’m I’m way ahead of everybody else. You know, some people would try to posture that they were ahead, but the the reality of it is everyone seemed like it was a something they knew they had to do, but they didn’t really have a g a good playbook. And so I started you know.

thinking about that and taking all the feedback that I was getting from the executive teams and CEOs of the companies I work with. And I started you know, pushing them into a a framework. And I said, you know, if we had a a simple framework that would give people a roadmap, I think that would that’d be really helpful.

Rick (03:11.639)
Is it is it fair to say that there’s a lot of information out there with AI and it’s not necessarily tailored to the growth elevated audience and maybe you’ve shortcutted that with this with this approach?

Julian Castelli (03:23.696)
It’s absolutely true that there’s a ton of information. I think there’s over information overload. I probably listen to three to five podcasts a week and I’m getting bombarded with this, and it just makes me feel anxious that I’m not doing enough. But it doesn’t necessarily tell me what to do next. And the approach I took was rather than you know, and then the other thing I hear is, you know, you know, go do this with Claude this or or go do this with Gemini. And, you know, you you dive into, you know.

Rick (03:37.037)
Mm-hmm.

Julian Castelli (03:51.255)
which which models and like does that really matter? I I I was thinking about the people I work with every day. They’re the CEOs of their companies, they’re leading a group of people, they’re they’re they’re maxed out, they’re busy, they know they have to use AI, but but they’ve got a bunch of competing priorities. And in fact, that’s that’s that’s the opening parable of of stage one, which is about raiding a bicycle, which I’ll I’ll talk to you about. But they they

Rick (03:54.307)
He

Julian Castelli (04:18.8)
They know they should ride the bicycle. The bicycle’s AI, but they they don’t have time. So so they carry the bicycle on their back while they’re trying to run. And and you know, this was actually one of my CEOs. I was talking to him. It was probably the fourth time I brought up like so, you know, we’re an AI company. This is a company that is its product is AI, but it was it was it was formed, you know, before you know eight years ago. And

It it it has AI in its product in terms of compute computer vision, but its go-to-market process is pretty straightforward. It’s not you it doesn’t use a lot of AI. It’s its RD process doesn’t use a ton of AI, it’s it’s its hiring process doesn’t use a ton of AI. And so I kept talking to him, so hey, you how are we gonna move to be an AI first pro company? And he was like, you know, Julian, but you know, I I I totally agree with you. I haven’t progressed on that initiative. You know, is it more important than closing the next few customers? I said, Well

You know, it it it it it’s not not on its own, but but it it that that’s the dilemma. And then he’s like, Is it more important than hiring our head of sales? Is it more important of the you know than doing this and raising our funds? And and as he was talking, he stopped. He said, this is this is like the parable of the man on the bicycle. And and I said, What are you t what are you talking about? And he said, There’s this old Indian parable and it has this this man who’s running through the desert.

In fact, I’ll I’ll show you the picture here.

And he’s running through the desert and he’s he’s anxious and he’s he’s he’s clearly late and he’s he he’s he’s sweating, it’s hot. And there’s a there’s a villager on the side of the road, he says, My friend, why are you running? And and you know, you have a bicycle, why don’t you ride? And the man ri running with a bicycle says, I don’t have time to learn, I’m already behind. And you know, that was such a brilliant analogy because you know, he’s like, That’s

Julian Castelli (06:16.492)
I’m I’m making excuses not to learn how to use AI or or to implement it with my company because I’m too busy. Yet AI could be the solution. Right. And this is a I’d say it’s a traditional company. It’s it’s a SaaS company, it’s a company that has AI in its product, but it’s not AI native. It’s a company that is that started eight years ago, and so AI has is kind of been born midway through the company’s processes and it it it

It is just now adopting it. And this is the same as many of the companies I work with. They’ve been around for five to eight years. AI is an incredibly new tool. and they’re all using bits and pieces of it, but not one of them was could just say, yeah, we just switched and overnight we’re AI native and we’re doing five million dollars per employee. It’s great. You just you hear these things about doing five or ten million dollars per employee, or there’s two people companies, and you’re just like, I I I I don’t know how to go from here to there. Right? It’s just like where do you go from that?

Rick (07:06.018)
Ha ha ha.

Rick (07:13.292)
Yep.

Julian Castelli (07:15.328)
So long-winded way of of answering it, I started you know, talking to the CEOs about about this, talking about the executive teams, and I said, All right, well, I’m not an AI expert. I’m not gonna tell you about Opus this or Claude, you know, you know, Fable that. You know, I’m I’m not I’m not the expert on the AI. I’m trying to learn like everybody else, and I’ve got my own FOMO, I’m carrying my own bicycle. But what I I am good at is is helping executives prioritize and do change management.

and and and lead companies through through change and that’s really what this is. So I put together this three stage framework that I thought would be simple enough that can give company leaders a roadmap of how to get on the bicycle and start using AI across across three stages and I’ll show you those those three stages. The first stage what we’re gonna talk about today and it’s it’s called Knowledge Worker

AI enablement. So anybody who’s sitting in front of a computer, you know, you want them all riding the bicycle, you want them going as fast as possible, you want them all learning. But if the CEO is too busy to figure out AI, well, what about the person who’s working 60 eight hours a week, you know, on their deadlines or trying to close their sales or try, you know, has a bunch of customers to help and customer success, they’ve got their own bicycle challenge. And so this is built for the CEO. It’s a process for how they can enable all of their team members.

To you as AI, and that’s what we’re gonna talk about today. Stage two, which we just went through in a workshop in the community the other day, is AI process transformation. And the idea is that once you’ve kind of built general knowledge and practices and expertise within your organization, you’re far better able to actually attack core processes and start automating them. And that’s the process of stage two. And this is where the real value comes. You can save costs.

You can accelerate time to revenue, you can improve win rate, all kinds of things you can do. This is the the the sexy stuff, but what I found is companies that just jump there struggle because they don’t have the base of of AI knowledge and and comfort. So we recommend that we do it in stages. That’s stage two. And then stage three we’re gonna do in a few weeks, and that’s how do you put AI into your product so that you can not only be operating with AI across your company, but your product.

Julian Castelli (09:36.935)
is a tool that might help other companies accelerate their I I I journey journey. And that may be the most complicated one. I’m looking forward to getting the feedback from from the community as to some of the ideas we have in there. But today I’m happy to walk you through stage one.

Rick (09:51.01)
Awesome. Yeah, this is great. do I I assume you’ll get to this, but do you have a strong opinion about whether AI enablement and adoption needs to be top down you know, through like kind of leading in front of the pack versus like, you know, you do this.

Julian Castelli (10:10.308)
I I d I do. i i i I call it the CEO mandate. and I’ve seen this with companies I work with, with peer companies, that if this is handed off to someone else, you know, there’s a there’s a good good use for a chief of staff or a COO here to be the champion. But if it’s handed off to someone else and then the CEO continues business as usual, it’s not gonna work. and and you’ll see that you know it’s it’s it’s it we talk about the CEO has to lead from the front.

And the companies I’ve seen do the best at this is when the CEO dives in heads head first and just becomes immersed in AI. So they have to be to to to beat our analogy to death, they have to be on the bicycle out front showing everybody that they know how to ride a bicycle and they’re riding a bicycle every day and and and there’s you know there there’s there’s almost no no use case where they wouldn’t. And I think it’s the same, I’m sure you on your journey, the more you use it, the more you see the use cases. So it really requires everyone to jump in the pool or

R jump on the bike, whatever your your analogy is, and that people follow their leaders. So if the leader’s like, okay, hey, take an AI pill and we’ll see you next week, so I can check that off the box, you know, everyone’s gonna everyone’s gonna, you know, do as I do, not as I say, right?

Rick (11:21.667)
Yep, a hundred percent.

Julian Castelli (11:23.891)
So I think that’s a really important point, Rick. so yeah, let me let me walk you through through through the stage one here.

So I’m gonna walk you through the the idea is that this is supposed to be simple. The idea is that this is something that could be done in 30 days, and it it it’s fairly straightforward. So the the the the first step is to appoint an AI champion. The second step is to publish an AI charter, and we’ll go through each of these. The next step you can do by week two is to provision the tools.

Also in week two, you can build working groups. Then you once you once that’s all set, that’s where the CEO has to both deliver the mandate and start getting out front and and being the champion of of this whole thing. They’re gonna have an AI champion who’s kind of the organizer. It’s kind of like the the implementer, but the CEO has to be the champ the the the champion and and really own own the results and and celebrate the results.

CEO deliver mandate means gets everyone working on AI on their own individual workflows. And then we launch work working groups. And this is an idea that I’m super excited about. It could be once a week or twice a week. I recommend twice. And then we we we we you’ll see how that works. And you you set a showcase date for 30 days out and you start having a monthly showcase where where maybe it’s your all hands meeting and you start providing

a a shared celebration of AI wins and it becomes part of your your all hands meeting where everyone is seeing who’s do using AI and what use cases they’re using and how they’re being creating creating value within the company.

Rick (13:20.525)
Yeah, it’s awesome. have you run into any any pushback from people on tool provisioning? I I I know that some CEOs can be I mean I I can s I am scared sometimes to give I don’t know if you’ve seen the this is not necessarily a appropriate example, but have you seen the the gifts where they give the monkey a machine gun and they don’t know what like they start shooting everyone by accident because they don’t know what’s going on? That analogy works when you give some

Julian Castelli (13:42.767)
Mm-hmm.

Rick (13:49.762)
you know, the wrong person AI and with with the wrong security protocols. Like I s I I imagine there’s a lot of fear out there on on who to give what access to. Is this are you are you a champion of like, hey, this should be you gotta kinda err on the side of more access versus less?

Julian Castelli (13:54.925)
Yeah.

Julian Castelli (14:08.493)
Right. So so the the the way I address that is there are you you you need to have a charter and you and your charter is gonna incorporate the specific tools that you’re gonna use. So all the issues that you’re talking about are very real. And frankly, this is when you listen to all the podcasts, you like it it’s changing weekly in terms of which tool to use and and how much cost and and and and what’s safe and what’s not. But the point is as a leader, your job is to say

Okay, this is what we’re gonna do. We’re gonna go with chat GPT. Here is, you know, here are the rules. Here’s the charter. In fact, let me move to the charter because that that’s that that talks about answering those exact questions, right? And there are more complex charters that you might find. I try to keep it simple, but basically, you know, as you you work with your AI champion and if you need a third party resources, there are plenty out there, but don’t expect the your employees to figure this all out, right?

Make a decision, figure it out, say we’re gonna use Claude, or we’re gonna use Gemini, or we’re gonna use Chat GPT, and go through all the questions that you just asked. What what plan? You know, who pays for it? Is it company reimbursed? What’s our policy? Right? Let’s create some rules. What data should go into the tool? What data should not come into the tool? Okay. Is using is there an excuse called, well, that was my what my AI produced? Is that an excuse?

The answer should be no. Just like it you can’t say, well, it that’s what the spreadsheet produced, or that’s what the cell phone produced. These are all just tools to do your job. So you have to train your people that the human owns the output, but the AI is a tool to to do to accelerate the output or or to structure your output or to enhance your output. So you have to really set down the rules and make sure that like

Hey, you can get fired for using AI. And probably people will get fired for using AI to your point. If if someone picks up a machine gun and sprays it around like a monkey and hurts somebody, they’re gonna get fired because that someone is our employee. And it’s their responsibility to learn how to use the the the tool, right?

Rick (16:16.567)
Yeah, so I I see one page here, in the middle of six rules and everyone signs. Is that is w why why one page?

Julian Castelli (16:31.595)
You know, the the idea was it to be simple because I’ve I’ve talked to companies about this and I’ve said, Well you why when are you gonna start? And they say, yeah, that charter thing. I think I’m really gonna do that. I’m gonna talk to some some consulting agencies and you know, I talked to them sixty days later, like, How’s it going? yeah, well we’ve scheduled the first meeting and we’re gonna have another meeting there, then we’re gonna do an audit and they spend, you know, a hundred and twenty days farting around. I said, Look, you can do this in two days. Get together with your core team.

Create your charter. Here are the critical things you need. And then every six months revisit this. So if there’s something you left out, put it back in, put it in later. Make it a working document. But do not let creating a charter become a project that you push out progress for a quarter or two and and wait. Because that every moment you’re waiting, your whole team is running around carrying bicycles. Or

There you know, some are riding tricycles, some are running skateboards, everyone’s doing different things and you might get into trouble because you don’t you haven’t provided leadership. Your job as a leader is to provide leadership and say, This is what we’re gonna do. If we find out that, you know, Gemini was our choice and there’s a problem with Gemini, in six months we might change it. We might go to Claude or we might go to ChatGPT or we might go to Groc. Things are changing rapidly. Your employees need leadership. The idea of one page was make it simple, get it done tomorrow. Like you should have the first draft within

24 hours. And then you can go out and share it with people and get some feedback and you can refine. But just the fact that you know these categories, there are six categories here. Maybe there’s eight that you want to use. Or maybe there’s maybe you decide on four. But the the fact that you actually get together and and create one, I recommend six to start, means that you probably know the two others that you might add because you’ve thought about the problem and you’ve forced it on one piece of paper.

Rick (18:18.051)
Yep. Makes sense.

Julian Castelli (18:21.267)
So I think this is something that that you know is a good kick start and and a lot of companies a lot of CELs I thought like, I guess you you know I guess that’s right. We’ve done kind of two of these, but the other ones we only think about, and it’s in this gray area. So there’s a lot of content out there about the perfect charter. There’s a lot of companies you can hire to help you create a charter. You know, my point is just create one and and make it a living document. So quickly put a charter together.

The next thing to do is name your AI champion. And this is a an important distinction I want to make. This is not your IT guy. This is not your CTO necessarily. I would say that this is your implementer or your your integrator in the EOS framework. It’s it could be your chief of staff, it could be your COO, could be your CFO, it’s probably not your CEO. It’s someone who’s gonna be doing cross functional leadership and has authority.

Within the company and is going to make sure that the trains run on time and it’s going to take this project, which is really a cross-functional educational project, and run that project. That’s who you want this person to be. And guess what? You yeah, the first job for your AI champion is like, hey, go research 25 AI charters and come back with what do you think the best one is for us, right? Using this framework. And so they can they can help you create the the charter. but it’s not a committee, it’s not optional.

Rick (19:38.115)
percent.

Julian Castelli (19:44.121)
It’s a it’s a very important job and you know, frankly, I think it’s a very attractive job. If if you know, I was asked to be an AI champion today, I think that would be a pretty awesome thing to put on my resume given you know what’s what’s happening right now. And to lead a company through AI transformation should be a very attractive experience for any anyone in their professional career.

Rick (20:03.641)
Totally. I I mean I I would imagine that this is what if if if if the person that you want to give this to is not already wanting to do this, they’re not the right person.

Julian Castelli (20:12.73)
Yeah, exactly. So you’d want to find someone who has some inclination, but also has the cross-functional authority. Obviously they have to be organized and and like it says it doesn’t have to be any type of title to start with, but it has to be someone who’s who’s gonna run this program. And the key is the CEO still has to be the champion. You cannot hand it off to the to the AI champion and just say, Okay, it’s all up to you. Just like let follow Rick and see what Rick does. The CEO has to be the co champion and has to, you know, lead by example and be

Rick (20:15.008)
Mm-hmm.

Julian Castelli (20:42.032)
you know, at every race and and and and participating. but but that you don’t you don’t want your CEO scheduling working groups and and and tallying votes and doing all the administrative work that this is g this job is gonna require.

Julian Castelli (20:58.617)
So when you have your champion, now you’ve got a partner to help you implement this program. Okay. And then this is when the CEO mandate comes in. And so the two things the CEO has to do is one is they have to participate and lead from the front. And two is they have to create a mandate. And this is the one where I’ve gotten some pushbacks from CEO. And this is the this is literally the bicycle analogy coming to bear where you’re like, my God, Julian, really? That’s a lot of time. You want, you want two out two hours per.

per today from every employee? And I said, yeah, well, you know, how how many how many hours you your employees working? And you know, some say 10, some say eight, you know, whatever it might be. And I said, well, how about this? What if you split it and say, take one hour from the company, and then you you work an extra hour. Whatever you’re working, you’re gonna work one extra hour until you’re fluent with AI, because it’s gonna probably give you six hours back once you start using it. But for the next 90, 120

you know, maybe a little bit longer, days, you know, we’re gonna ask you to work an extra hour and that that that could be at night, could be in the morning, it could be during your work day. And then, you know, you can take one hour from us when you work you’re you use AIs too, but you should be using it twice, two hours per day at a minimum. And the the idea is to get everybody thinking in that like if they’re not touching AI two hours per day, they’re doing something wrong. And this has to be kind of harsh.

on the most extreme side, I’ve talked to people that have put in the measurement devices where they literally can tell how many hours people are using AI, and that sounds a little bit brig brother. But you know what, they found that it’s been very useful. And and and and you know, that that company example, I in fact I had Kathy on the podcast earlier, she said that you know people were blowing the two hour mandate away. They were they were working at four or five, six hours, which is once you start riding a bicycle and you’re f you can do it, you can ride with no hands.

You start using it, but if you don’t, you’re still carrying the bicycle, right? If you’re just trying to be give it lip service. so the idea is that you make a temporary investment and you yeah, you might have some inefficiencies for two or three months, but if you do that and you become fluent in the language of AI, you’re gonna be riding so much faster and you’re gonna those hours are gonna come back to you.

Rick (22:57.337)
Mm-hmm.

Rick (23:17.123)
How long does this typical mandate run? Is it is it a couple of weeks? is it a quarter? Or is it just until until AI is fluent?

Julian Castelli (23:27.029)
I think it’s still until AI is fluent, my guess is it’s gonna take three or four months. And you know, I’m leaning towards some sort of measurement. I at first I thought it was you know, I think it was about a year and a half when I did that podcast with with Kathy. And now I’m at the point where I think measurement is worthwhile. Because if if people you know, if you can tr truly see that people aren’t using it, then I think you’re is it’s telling you something about the buy in that you’re creating in your organization.

And you’re also gonna be able to see who’s using it the most. Now, the the funny thing is when I presented this to the to the growth elevated and YPO groups last month, people were talking about the token maxing. So there’s some places now where this goes the opposite direction. So instead of two hours, they’re using 24 hours of AI and and they’re ranking up huge bills, right? So you you could very quickly go from like, my god, I can barely get people to use

the AI to my god I I’ve just spent a million dollars on AI tokens, right? So that’s part of the the charter and yet y you and your champion need to do is is create some some bounds and some limits, right? And you know, don’t just set the AI on some useless task to try to, you know, have a vanity project, but you know, y use it. And and and and to some extent, you know, there’s definitely some validity to say, okay, how many tokens are people using? And if you can see the output and see what they’re creating, there’s you know, I I I

Candidly, if I was in a seat right now, I’d like to see the the output or the token usage per employee because it probably would show me some measure of pro productivity. Now that being said, it’s not a set it and forget it thing. Like I said, you might just get people, if you just, you know, tri take one metric, you might get people gaming the system, which is what what those stories about out of Silicon Valley were. But the other stories about Silicon Valley were the people who truly used AI, you could, you could recruit people by sh by talking about your token budget.

Right? People are saying, hey, you’re getting fifty percent of your salary in tokens, token dollars, or a hundred percent of your salary. You know, we’re gonna this is a three hundred thousand dollar job and you get three hundred thousand dollars of tokens. That matters to people who really use aid. That that’s the positive that you want, not the people who kinda, you know, concoct some AI solution that just kinda grinds all night so that they’re on the leaderboard, right? Yeah.

Rick (25:32.279)
Yeah, yeah. You don’t want the person who you have to teach AI. You know, you want the person who are is AI fluent and then you want them to you want to be able to attract the best one. D does what one thing that I’ve I’ve run into is when you first get someone on on Claude like and they’ve got some AI experience, the first question they ask is, What happens if I hit my limit? and and w what I’ve learned from from administrative roles in this is that you can set you can govern this through the tool. Like it’s not something you should

Julian Castelli (25:35.134)
Yeah.

Julian Castelli (25:59.103)
Mm-hmm.

Rick (26:01.25)
You you don’t have to you should write it into the charter, but you don’t you can use the tool to manage this. but but my my message has been I I’m curious if you agree with this is please just max out the tokens and then let’s talk about it. I don’t care about maxing out tokens until you actually do it.

Julian Castelli (26:05.78)
Mm-hmm.

Julian Castelli (26:13.15)
Yeah.

I call it the pajama moment. I I don’t know if if you’ve had the same one, but I’ve been sitting in this chair late at night or early in the morning because I’ve had to like, you know, I’ve set Claude I’ve I’ve done all the work to I I don’t know if you saw my thing about the ETL monkey, but I’ve I’ve I’ve got all my stuff in the right place. I’ve pointed Claude at the right information. I’ve briefed Claude, I’ve gone and done all this stuff, and it starts working. And I’m sitting there and I’m like, you know, you know, it’s i it’s either I’m gonna wake up tomorrow morning and and go do it, I’m gonna be, you my pajamas, or I’m gonna, you know, be up to one in the morning waiting for it to do it.

But that’s when you know it’s it’s kind of like you know riding a tent speed for the first time. It’s like, actually I’m really I’m really stressing the system here. I’m I’m pushing the limits to what it can do. And I think we all went through that time when when we all kind of got upgraded to the next plan because we were, you know, we were using it enough. And the funny thing is I I didn’t stay it there. Like now I’m back to kind of a normal thing and sometimes I use it, sometimes I have to go up to there. But that’s part of the thing that everybody has to learn. I I that was me learning how to use the tool. And I’ve actually created some

some efficiency tools by you know creating markdown files periodically and so it doesn’t kind of have to learn everything every time, right? And that’s that that’s mentally you you start understanding that just the way like you know you go to school where or you go to the library, that’s where the information used to be, right? And you go and you spend time learning, pulling information so you can solve problems, right? That’s the same sort of thing you do with with the ETL. You extract information, you transform it, you load it in a place where the model can have access to it.

It actually, you know, walks through, you know, hu hu human neuroscience pretty pr pretty neatly in a neat neat way. But the truth is you have to do it yourself to really understand. It’s not something you’re gonna just learn in a course that that easily. But I think that’s that’s part of the charter, which is, you know, okay, when you r run out of tokens or you you hit your max, can you use the pro plan? Who can use the pro plan? How are you gonna decide that? That’s something for the the AI champion to to to determine as part of the charter. And that’s something you might change over time.

Rick (27:47.461)
Yeah.

Julian Castelli (28:12.643)
Of course now, you know, if you listen to now do we do it internally versus using the you know, the hyperscaler models every week there’s a new debate, right? So th this is the kind of topic that you need to find someone who’s gonna be listening to all that and getting all the best practices and bringing that into your organization.

Rick (28:12.921)
I like this.

Rick (28:28.377)
How d like let’s say you do all the right things to this point, you deliver the mandate. I’m assuming it doesn’t just happen overnight. Like what what what do you what should people ex what should a CEO expect once they deliver the mandate?

Julian Castelli (28:40.696)
Well, this this is a good good s sequence the next page, right? Everyone’s gonna say, Of course we wanna use AI. absolutely, but you know, after I close this big deal, or after we recruit this person, or after we we close the books, or we’ve got to close the quarter, or we gotta we gotta raise this round, and then we’re gonna do AI, right? Because everyone you get pushback and and and often it’s from your best people, right? See it’s the the person who’s the most busy who has the highest quota. And and and that’s because they’re they’re running.

And you know, the example I again I use is they’re running with the bicycle. and you know, they’re they’re too busy running to accomplish the goals the old way, and they haven’t figured out how the A AI can help them solve those same problems. And so that’s where you have to be very careful because if you let people skip this for too long, then you’re gonna have erosion both from the bottom, just like you’d have it from the top, if the CEO’s not doing it. So this is the big

constraint that I think you have to overcome is you’ve got to get everyone doing it. And I think you know, it it’s it it’s it’s some friction, but I think it’s temporary friction. you know, because the the busyness that the that whatever it is, whether it’s closing the deal or recruiting or or coding, whatever it is, once you learn how to use AI, it you know, it’s gonna look back like, I’m too busy using my slide roll.

You know, let me finish doing this. I’ll I’ll learn how to use that calculator later. I’ll use Excel. you know, I don’t have time to learn how to use Excel. I w I’ve got to get my work done using the slide rules. Like, okay, take some time and learn how to use Excel. Right? It’s the same thing. It’s the same work that’s being created. It’s just it’s just a tool that allows you to do it faster, better, you know, more more in more detail.

Rick (30:26.489)
So this it’s kinda like the CEO should expect the same resistance from every work some workers that they’re probably experiencing themselves before they adopt this approach. Yeah.

Julian Castelli (30:34.976)
Well, that’s that was why the the bicycle theory was great, because they’re feeling it. It’s it’s like the whole FOMO thing. I know I should be doing it, but I don’t have time. I’ll do it after I get this done. And if the CEO’s doing it, you can get bet that you know, not 100%, but maybe 70, 80 percent of the employees are probably doing the same thing. And then there’s the twenty, thirty percent that are curious, right? And so this is how we start leveraging that that dichotomy, right? We we create working groups because there are some people

Who are just AI curious and they’re the early adopters, right? There’s there’s an early adoption curve in every technology. And the idea is, you know, how do we learn? We learn that pajama moment that we talked about. We also learn by talking to our peers and getting experience. Like, you were able to do that? How how did you do that? Can you show me, right? And so the idea here, and this is the most important part of the whole program, this is where the learning works, is we want to create working groups. And I think this is good not only to to drive AI.

sharing and adoption, but also actually creates camaraderie amongst the team, especially people who are working remote. So you create two one hour meetings a week and the champion puts people together to create groups, five to ten people per group. I recommend that one group’s departmental so it’s it’s it’s your people you work with every day and you’re just talking about AI projects. And then the other one’s cross-functional. So if you’re in finance, you might be talking to someone from tech or someone from sales or someone from HR and you’re getting to know people and you’re sharing these use cases.

in a way that is is is super valuable because I’d say eighty percent of my big leaps have been c have come from someone saying, have you tried Claude Skill or you try I got more FOMO. I gotta try to figure out what Claude Skill is tonight, right? And I go home and I, you know, and I try to figure it out myself and then I meet them I see him again. Hey, can you tell me like again what you used it for? Like, I created, you know, this marketing guy

Brand book, and all of a sudden I just apply the brand book whenever I need to. I’m like, okay, I see what you’re talking about now, right? And so the example for these working groups is is very structured. it’s basically a one hour per week, it’s a one-hour meeting. The champion needs to create a moderator for each group. The it’s a very structured agenda.

Julian Castelli (32:50.825)
And you basically have a the a little intro, then you have three 15-minute presentations, and then you score the presentations. And so it could be as simple as, you know, hey Rick, you’re you’re up next week. So you know, you’re gonna be sharing your AI use case. And you know, if I’m the moderator, I’m gonna pick three people and I’m gonna moderate that meeting and we’re gonna get those use cases at the end of it. It might be that you automated an email. That’s fine. It might be that you automated a memo, it might be that you just, you know, this is individual knowledge worker tasks, not

We’re not splitting the atom or or or automating workflows. We’re basically just, you know, the stuff that you do every day, going through your email, doing research, looking things up on the internet. I mean, to this day, I’m I’m having conversations with friends, like, my God, Julian, I just did this. I just had used AI to use this. That’s that’s that stage. You’re discovering what you can do. And by sharing with a group, every week you’re gonna hear six use cases and they’re gonna probably be different just because of, you know.

Rick (33:37.034)
Mm-hmm. It

Julian Castelli (33:46.951)
peer pressure, you’re not gonna want to present the same one that Rick did last week if I, you know, I’m gonna try to cut to something new, right? And so everyone’s gonna share these use cases. Some might be overlap. And then at the end of it, you just vote like, okay, on a on a scale of, you know, ease, you know, ease to implement and value achievement, how are we gonna rank these? And so the moderator scores their ever every use case. And you can see if you do this for six to eight weeks, you’ve got a eight

Eight times 24 presentations, actually eight times six. So you got 48 presentations, all of a sudden everyone’s seen 48 use cases in in eight weeks. That’s pretty good. and everyone’s gonna naturally have that peer pressure that they have to present every two or three weeks. It’s not a lot, it’s 15 minutes. It could be you there’s there’s no bar, doesn’t it? You you don’t have to be splitting the atom. You can literally just say, Hey, I was just playing around and I did this, or I I I shopped for that, or I did a research for this.

Rick (34:27.211)
Is is is the I

Julian Castelli (34:43.526)
What you’re gonna find is the natural peer pressures, you’re gonna go from the basic knowledge worker, you know, you know, look something up in my calendar to actually I thought, why can’t we automate this or that? And and and the the the value is the whole tide is going to rise as everyone starts using AI.

Rick (35:03.171)
The the thing that that I’m thinking about is how do you actually make this happen? I think if one, you’re asking people to spend two two time two hours a week on or excuse me, a day for some amount of time. Does this like does this count to one of those this counts, okay.

Julian Castelli (35:18.46)
This counts. This counts. So so if you have two meetings, so you so we’re saying, you know, you have to use AI for ten hours a week. This is two of the A eight, two of the ten.

Rick (35:27.373)
Got it. Got it. And and in in and in this case, every single person in the company is gonna have two groups, one cross functional group and one department group.

Julian Castelli (35:35.149)
If you’re if the company’s large enough, that’s what I recommend. I I’m thinking of the a hundred plus people. If you’re thirty people, maybe there’s only one group, right? It may be a little too much. But two groups really gives you the benefit of the cross functional. So someone in finance is going to be using it differently than someone, you know, in HR and and so or someone in legal. But I think the the the exploration and the learning really is is exponential when you start going to different functions. So if you had to choose just one, I would choose group two.

Rick (35:41.783)
Yep. Yep.

Rick (36:01.646)
Yep, I agree.

Julian Castelli (36:05.048)
Not group one. I would I would do cross functional.

Rick (36:07.619)
Great. How do you like th the other thing I’m thinking about is, you know, this this assumes I’m assuming every company has some form of an operating cadence. When you when you put new meetings on the on the calendar each week, without maybe getting rid of something else, it can cr cause problems for people. How how how do you think through those trade-offs? Like what do you take off to make room for this?

Julian Castelli (36:32.077)
Yeah, I think I think that’s probably a a good reality check. That’s the job of the the AI champion. That’s why you need an integrator type of person to kind of evaluate those roles, look at the you know, who understands the company cross-functionally. And yeah, I I think that, you know, it it could be a good idea to say, hey, for for the next six months, we’re gonna eliminate XYZ and make this a priority. I think that that symbolically shows that you’re making hard choices. You can’t do everything. Right? So it might be worth l eliminating something to

to put this in there. but it you know, when they are and how you fit them and whether they’re in person and and how you organize, that’s the job of the champion. It’s not what you want your CEO obsessing over. Maybe the CEO approves it, but you want that champion to set this all up and and and tackle those problems. This could be different for every company.

Rick (37:18.733)
Yep, that makes sense.

Julian Castelli (37:21.336)
But that’s why it’s a it’s a real role. It’s not easy, right? Okay. And then what what what the champion needs to do is they’re gonna get at the end of every week, they’re gonna get the log of all the wins, right? And you know, some of them are gonna be just basic, you know, ba basic like Excel spreadsheets or or emails or or word documents. It’s like, okay, they’re gonna fall to the bottom, but eventually you’re gonna start having in the rankings, you’re gonna have some pretty cool automations that are just kind of rising from the bottom, right? So you’re gonna log and and stack rank.

based on the the the scoring, you know, on impact and ease, all of the the ideas. And and some of these might actually be, well, I only got a third of the way through it, but this is a really cool idea. And the idea is that, you know, you you feed the groups and say, hey, even if you only got halfway through something, you think it’s good for the company, you know, enter it as your project because it might be something that we we do, we, we sponsor as a company, right? So what you’re doing here is you’re building your backlog.

The other thing you’re doing is when you when when someone actually creates a really good let’s just let’s say they do an interviewing skill or a marketing skill, hey, this is something that could be a company asset. So the other thing you do is start creating a library so people can share, not everyone’s rolling their own every single time, but like, hey, we have an official skill for our marketing presentations and we have a you know a skill for recruiting and we have, you know, basic things that now this organic activity is creating company assets, be they projects, prompts.

skills, you n you know, you name it, custom GPTs. And so you wanna you want to harvest those from the group and and make sure that they’re they’re utilized. And now one person builds something cool, everyone can use it. Now you’re starting to get a little ROI on that investment in addition to just your knowledge workers, you know, riding the bicycle at at a higher pace. Okay, the last thing is you you know at the end of the quarter,

You know, you’re that champion or that becomes a group now is gonna be saying, okay, we’re gonna celebrate the top three projects. And so you you you share the best use cases, you present them, maybe there’s a little a little show and tell. you know, they’re gonna they’re gonna be some pretty cool things. you know, maybe the company maybe you can make it a contest where the company votes real time on the three that are there, so you can you can have some fun with it, you can award prizes, you can you can identify

Julian Castelli (39:38.281)
Champions. I mean, the real cool thing is you’re you’re creating buy-in, people are learning, you’re kind of validating this is important, that this is, you know, you’re investing in the future of the company, and you’re recognizing your your your future stars, like who are the AI native or the AI curious people that really want to lead here? They’re gonna be the ones on the leaderboard, and they’re gonna be the ones that you’re gonna need for stage two. And that’s that’s where this whole process takes us, is we’re we’re actually, you know, building the setup for stage two, because

Now we’ve we’ve enabled our our workers. It’s running. People are doing things with AI. It’s no longer a chore to go to these meetings, probably looking forward to it because they it’s just the way they work, right? They’ve gotten that they’ve gone from nine hours, now they’ve got the equivalent of, you know, maybe five hours of busy work, but they’re getting, you know, much more work done with using AI. Your process transformation, you’ve got your top list of of all these winners that, you know, maybe they were done like at a individual desk scale, but now could we do this?

at a company scale that you know is is is is a company process. So when we talk about stage two and we start creating company sponsored projects, we’ve got fifty, a hundred ideas that have percolated bottoms up, you know, as as the things that we might be able to work on, which is hugely exciting. And then, you know, once we do that and we and we’ve been successful with that and we’ve we’ve we’ve both trial and error, not always success. Sometimes you you learn from failure more than success. Well now we’re gonna we’re we’re ready to start

Saying, how do we put this into our pro on our product so our product becomes solving other companies’ AI problems with it with AI, right? And that’s that that’s stage two and stage three. But your working groups are your RD engine that you know gives you your ideas, your innovation log is your opportunity backlog, and then your showcase is your prioritization meeting. And so it pulls you right into stage two. And and the idea is that this is supposed to be fairly straightforward and easy, not too technical.

But at the end of this, to to where we started talking about, you know, do I use this model? Do I use that model? You know, can you know, is is someone gonna get in trouble for doing the wrong thing? You could see how a company that’s gone through this process, even for three months, probably has solved most of those issues at this point.

Rick (41:49.939)
What so you think like a typical timeline to get through stage one and s and sort of be ready for stage two is what, somewhere between forty five and ninety days, or is that too aggressive?

Julian Castelli (42:01.222)
I think that that’s the early part. I think it’s probably three to six months because I think reality is it’s gonna take thirty to forty-five days to launch it. even though you could do it as early as quickly as a week, right? you know, it’s better than what I described earlier. You spend two months, three months just get a charter. Right. but I you know, my guess is that the you give some people some grace. If you get this done in three to six months, I think that’s pretty good. And you can also overlap. So once once you see this is cooking with gas and you’ve got everything running.

Rick (42:16.663)
Nothing. Yeah.

Julian Castelli (42:30.214)
You can keep this going as long as it’s interesting. You can bring in some different things. You know, the the working groups don’t have to just be show and tell. You can bring in resources. you can interview resources that might help you in stage two, which when we go through stage two, we’ll talk about bringing third-party resources. You can interview them by saying, Why don’t you come in and give a presentation on you know, AI topic ABC. And then you have another one give it to give a give a presentation on t on on that and then

And that could be like the the the company meeting. Instead of going to our working groups this this week, for working group two, we’re gonna go and have this this webinar and we’re gonna start learning together about the the latest thing that pops out of of ch of open AI or or anthropic, right? So you can do education that continues and you can evolve the program once the show and tell has kind of you know d met its core goal of of making everyone ride the bicycle, you know, surfacing ideas bottoms up from the company and then actually getting people

used to experiment with playing and and building things. That’s really what you want to get to. At that point, look, the people who get it are going to start using it going forward. You’re not going to have to monitor. Like if you if you’re using one of those token tools, you’re probably not worrying about the low ones anymore. It’s like everyone’s at a certain it finds their equilibrium and is is now using the tool. And the the thing is the outliers who are really being productive, those are the people you’re going to choose to say, you know, hey Rick, you know

I you’ve been crushing it. You were you’re you’re number two and five on the leaderboard. We want you to sponsor one of these projects that we’re gonna do in stage two. And that opens up career opportunities. So you find those people in departments.

Rick (44:01.082)
That’s that’s yeah, I I s yeah, it’s it’s interesting how if you do stage one right it it leads you right into stage two with with the plan almost already baked.

Julian Castelli (44:11.332)
Right. You’ve got you yeah, you’ve got the lists, you’ve got the champion, th th the the stars that are really the the on the leaders and and then you know, what what when we we get to back together and talk about stage two, it’s really about the process of how do you determine what you do internally, how do you leverage third party resources, how you leverage off the shelf software, how do you make sure you get an ROI and be disciplined about your projects, how do you report them to your your board and your ownership? and and and that’s a whole process into itself. But you’re

You know, find a lot of companies jump right to stage two and sometimes even to stage three. And it be ends up being a isolated group doing an isolated project, poorly reported on, not well understood within the company. You know, some succeed, some fail. I think they’re gonna succeed and fail in every every scenario. But I think with foundation of stage one, your odd your odds of succeeding and and and your odds of choosing better projects and your

ability to have more resources to th the that can knowledgely attack the projects is is greater. So therefore your chance of success is greater.

Rick (45:16.427)
sense. If someone wants help working through these stages, like any any advice you’d give them, can they can are you available to help people think through this?

Julian Castelli (45:26.849)
Yeah, 100%. I I I’m available to to talk the people through the process and the stages. And then, you know, as when we get into stage two, we’re gonna introduce the concept of an expert service provider. You know, these are the AI experts, right? I’m sure you you you’ve seen them, but the the the folks that are out there that it’s a it’s a birth burgeoning industry of people that can come in and and build your workflow tools. And we’re gonna introduce you know, we’re we’re gonna introduce

that idea in stage two. And I also, you know, I I have at least a half dozen of those those firms that I’ve worked with and I’ve I’ve deployed within my portfolio that that I can I can refer people to if they need the AI expertise specifically.

Rick (46:09.995)
And what should people look forward to on the stage two version of this?

Julian Castelli (46:14.474)
Yeah, so stage two, we’re gonna we’re gonna basically talk about, okay, now we’ve we’ve completed stage one. How do we prioritize company sponsored projects? We we okay, we are going to automate that process. Maybe it’s we’re gonna automate our recruiting funnel, right? And and what we’re gonna do is figure out how do we determine what project that we wanna prioritize with company dollars. Then we’re gonna look at, you know, how are we gonna attack it in a way like, you know, we do we build it internally? Do we buy it off the shelf?

And or do we contract with someone to help us build it because we don’t have that expertise? And then the third piece is, you know, how do we track progress? How you know, how do we, you know, it’s it’s the it’s it’s the scientific method. We we expect that this project is going to reduce time to completion, therefore reduce cost, and provide this ROI. So you start out with a hypothesis, you state the hypothesis, and then you measure, we actually did better than we thought.

in terms of the results of the the the the experiment. But strangely, the cost didn’t go down. Uh-oh. You know, that’s the that’s the headlines you’re reading about right now. So being very clear on, you know, how you create a hypothesis, how you test it, and then how you report on it, both for your executive team and your management and your board, so that you can build up the credibility that, you know, even even the things that don’t work, you’re learning and you’re getting better and you’re on that that wheel

Rick (47:19.171)
Mm-hmm.

Julian Castelli (47:40.329)
to to be using AI to to create the gains that everyone’s so excited about that are available. It’s I think it’s unmistakable that there are opportunities out there. It’s also unmistakable that people are gonna, you know, have some false starts and and some disappointments and then you know the media plays up both, I think. But we’re gonna go walk through a structure that gives I think gives you a a strong framework for choosing what to do, figuring out how to do it, doing your your your scientific method and and

doing good reporting and and learning through the journey.

Rick (48:12.825)
Well thanks for taking the time to walk us through stage one, Julian. looking forward to stage two and any final thoughts before we sign off today?

Julian Castelli (48:21.02)
Yeah, my my final thought is stop carrying the bicycle. You know, get on the bicycle and start riding. And then and then start riding with your hands off, then start riding with your eyes closed, then start jumping off of curbs and enjoying it. And when you when you run out of tokens in the middle of the night, you’ve probably have mastered the bicycle.

Rick (48:38.233)
That’s the win. That’s the first milestone. We’ll if if if you want more, check out growth elevated dot com. We’ll be following up with stage two shortly.

Julian Castelli (48:39.869)
Ha ha.

Julian Castelli (48:47.07)
All right, thanks, Rick. Great Red spending time with you.

Rick (48:49.187)
Thank you.

Introduction & AI Framework (00:00:00)
Rick introduces Julian and the Growth Elevated AI adoption framework.

Why CEOs Must Lead AI (00:09:51)
Julian explains why AI adoption starts with executive leadership.

Building an AI-Ready Company (00:11:23)
The foundation for AI adoption through structure, tools, and working groups.

The Two-Hour AI Habit (00:20:58)
Why every employee should spend dedicated time using AI.

AI Working Groups (00:28:40)
How cross-functional collaboration accelerates AI learning.

Capturing AI Knowledge (00:37:21)
Turning prompts, workflows, and experiments into reusable assets.

Preparing for AI Transformation (00:44:11)
Moving from experimentation to company-wide AI implementation.

Closing & Key Takeaway (00:48:21)
Why leaders should start using AI now instead of waiting.