806 | Every CEO Needs to Hear This About AI (Before 2026) | Brett Schklar

Return on AI? That’s the new KPI. Every investment in AI should prove it. That’s Brett Schklar, the no BS architect behind AI First Leadership and the author of AI Without the BS. He’s a trusted voice that over 1,200 CEOs have turned to when they’re tired of hype and ready for results. As a founder of Grow Powered and a former agency owner himself, Brett didn’t just talk about AI. He used it to build a 30-person business with zero employees. Now he’s helping service-based teams ditch fear, fix what sucks, and lead with confidence in the age of AI. But how does he do it? I meet CEOs every week. I tell them, take a list of the, you know, three, five, ten things that just you hate to do. In this episode, you’ll learn why agency owners must lead AI adoption before someone else does. AI adoption, it starts with one big thing, setting it at the top. How to empower your teams and clients to test, learn, and grow.

Skin your knees in the sandbox. Treat it as if it’s your number one employee, your number one coach, your number one consultant, your number one data analyst. How to AIify your pains. If you’re mapping your AI roadmap, start by asking, what is it?

You mentioned something about AIifying the pain. So what specifically do you mean by that? We all have to do things that we hate to do in our job or in our day. And it’s things that suck, but we have to do them. And they’re probably repetitive, rudimentary tasks. So I like to tell people, you know, to get a little bit comfortable, to get, you know, adopting this in a way that really does add value is take a list of the, you know, three, five, ten things that just you hate to do. And you’re like, oh, I have to do this. This sucks. I have to do this every day. And see what kind of those tasks can be automated, or how AI can be the research tool, or it can be an analytics tool, or it can be drafting responses to things that you don’t want to respond to. So take those things that you think really suck in your day, and add a little AI and make those same things suck a little bit less. There’s a tool called Fixer, F-Y-X-E-R dot AI. And it isn’t a tool that plugs into all of my five different email accounts. And not only does it categorize all the emails in a way that keeps things focused in my inbox,

because I’m also a recovering marketing person. So I have my attention that goes everywhere all the time. It keeps me focused. It also automatically drafts every email, a drafts a response to every email that it thinks deserves a response. And it doesn’t send it, because one of my rules is trust, but verify the stuff that comes out of AI. But it drafts emails for me. So all I have to do is take a look at those emails. And even if I’m correcting a few little things, it learns from the things that I correct. Now I just look at my drafts and I see which emails are already being responded to in the draft folder, and it makes me sound nicer.

It makes me sound like I pay attention to all the emails or to everything in the email. It makes me sound nice and warm and caring. And because of that, it saves me easily 30 to 45 minutes a day. But it’s also helped me close deals, because I’m a service-based business as well. And for me, I have learned that speed is one of the greatest assets for closing business. And so that really helps me out a lot. But on the more strategic side, one of the things that I love doing is working with companies who may be involved in the RFP process. They’re a request for proposal. And when you talk to organizations that have sales teams, who it’s the bane of their existence to respond to these RFPs, but you build a GPT within chat GPT, or you build an AI agent within co-pilot, or a task within Gemini, and you actually get it to respond to the RFPs, and you tell it, you don’t want to just respond to the RFP. You want to basically win against all these certain competitors. You want the answers to be unbelievably compelling.

And you want to really focus on the strengths of your company and downplay the things that you know that might be weaknesses within your company. It writes it for you. And there’s companies where I’ve gone in to talk to the sales team. And when we talk about RFPs, you would think I just talked about the end of days. They just hate the idea. But we rewired everything where it was an AI first approach to responding to RFPs. And there’s this one company that was probably responding to about two or three each month, because it was a six-week or an eight-week process, to now they’re going after every single RFP they can possibly find, which is more like 30 to 40 a week that they’re responding to. Again, going from a couple of months to 30 to 40 a week now. So with your lens and the way that you’ve been looking at this and all the people that you’ve talked to, what do you think agencies should be focusing on when they’re having conversations with their clients? What AI? A good business goes where their customers are.

You want to be ridiculously easy to do business with. And by adding these outward AI systems in place through a platform like yours that can give them an AI chatbot on the front face of a website, so people can take action when instead of having to make phone calls or send an email. That’s very powerful. We are seeing a lot of success in better customer experience by using voice AI assistance as well to answer the phone and to get people what they’re looking for faster. Instead of having to deal with somebody who’s got 15 things going on, they’re answering the phone, they’re also the front desk person that’s handling all the humans that are coming in and out of the office, putting people on hold, all of those things. So the goal isn’t just about saving the time. It’s making your clients more human and it’s giving them more places and more ways to interact in a way that’s more comfortable to them, to you, and makes the overall experience even more positive so that you become even more ridiculously easy to do business with.

Tell me how do you think agencies should be measuring their return on AI? So for me, what I find is the best way to get it out of that science fair project or a techie tool thing is to look at how do you evaluate the performance of your AI initiatives, hence the return on AI. There’s six real ways or six real lenses you can use to look at how you measure your return on AI. It could be saving time. It could be reducing costs and eliminating extra costs It could be the speed at which you lift revenue or the amount that revenue is lifted. It could be in productivity overall so that everybody becomes more productive.

It could be accuracy as a big indicator of return on AI, especially in places where there’s a certain percentage of accuracy issues that might be plaguing or haunting companies. And then client retention is also a really great identifier of a return on AI. And that’s a great thing if your platform measures those things and you’re already doing return on AI for companies. It’s just maybe not looking at it specifically with the AI lens on it. But yeah, that’s great that your platform has those capabilities. When it comes to AI, how do you recommend that agencies should best position themselves if they want to be AI advisors?

Yeah, I think the biggest benefit that people are looking for from agencies is help me solve some problems. And that doesn’t mean that the agencies need to become technical gurus. In fact, just the opposite. I think that if agencies are using AI in their own business, they’re learning about this, they’re investing in their own time and energy to make AI work for them. They will easily be able to translate those learnings and those wins to their customers and their prospects. So the beauty of all of this is, I love to say, this ain’t rocket surgery. You don’t need to be a rocket surgeon in order to really understand and adopt AI. You just have to get comfortable to play with it, learn it to say, will this solve a problem for me and try it?

So the fear of AI can be a real culture blocker in any organization that is looking to really take advantage of the technology and build efficiencies throughout their organizations. So when you’re talking to a CEO for the first time, and while a lot of people are adopting it like you point out, there’s, we still have a lot of laggards out there. For the CEO that is the non-technical CEO, what is the first thing that you tell them? AI adoption starts with one big thing, setting the tone at the top. The biggest impact I have is the easiest thing I do, and that is asking CEOs if they’ve had an open and honest conversation about AI with their employees. Most of them still haven’t. And so what that translates into is employees are scared to talk about their use of AI because they feel like it is, that it’s going to be perceived as cheating if I’m using AI. Or they’re scared to talk about how much more efficient they’ve become by the use of generative AI because they don’t want to go back to a boss and say, hey, guess what? I found a way to reduce 20% of my time or 20% of my time requirement while increasing the effectiveness and the accuracy of things by another 30% to 40%.

What else should I be working on? So when the CEO sets that tone from the top that we’re very interested as a company in AI, and we’re looking for people who are really trying to leverage AI in a lot of different ways, that’s the biggest change. And I just see it in the organization. When I talk to employees that are fearful of talking about AI, especially to some external person like myself, when they start getting that comfort level of talking about AI, when I have spent time with the CEO, the innovation and the ideas and the ways to deploy it really, it just changes gears completely. One of the things that you talk about is setting up a safe place to play or having a playground for everyone to experiment with AI.

Tell us a little bit about your thoughts on that and why you think that it’s so critical to have something like that before you’re releasing customer-facing AI solutions. The way I like to frame it is think about that you’re building AI for production, the real world, the things that you use, the things you want to train with as much accurate information as you can. But then you also want to create the safe space and think of it as like a staging environment where you’re actually making the mistakes. You’re trying things and maybe going a little too far with your prompts or a little too far with your research, and it might not be giving you the right answers. For me, creating a sandbox that’s something that doesn’t necessarily connect with the production environment gives a safe space to play around, to mess around, to make mistakes, to push the envelope, and great ways to do it. Most companies are either going to be on a Microsoft office environment or a Google workspace environment, one of those two.

So you think about the production environment being co-pilot, which plugs into Microsoft or Gemini, which plugs into the Google workspace. But if you treat a chat GPT instance like a separate one or a separate perplexity instance, and you say, we’re going to use that as the way to play around, learn a lot about prompting, learn a lot about how it finds information, but we’re going to keep it separate from the production AI environment. You’re training the good stuff, so you’re eliminating the garbage in, garbage out that causes problems for a lot of companies. But you’re also creating that place where mistakes can and should happen, but it stays separate from all of the business and mission critical information. What does a sandbox look like to you? And what does a successful one look like? Yeah, that’s a great question. So let’s take an example of trying to create AI assistance for responding to customer inquiries. So for example, agencies, it may want to help their clients, whether it’s a dentist’s office or a doctor’s office or things like that.

They say, we want to try to figure out how to do a really good job with responding via chat on a website so it feels more human and interactive. There’s a lot of things you have to do. It’s not as auto-magic as the systems would like you to believe. There’s a lot of bugs to work out. There’s a lot of prompting of not only to tell it what to do, but just as important to tell it what not to do and creating the guardrails about here’s the things that you can talk about with customers or website visitors, and here’s the things you do not want to talk about. Also, here’s a great example of something that we started to do for clients is take the company core values. It might be customer first, or it might be transparency, or it might be give answers that people might not want to hear, but they need to hear. Whatever those core values or companies are, build those in to the system that you’re using. The problem is, when you’re trying a lot of these things out, you’re making the errors. You forget to put certain guardrails in place, or you need to think about it with a 360-degree view instead of looking at it as far as telling it just what to do.

Skin your knees in the sandbox for a customer service example of customer and beta it, pilot it with a couple of friendly customers and get them to try it. You’re going to learn things that are working and things that are not working. Then once you work out some of those bugs and you stop skinning your knees as you’re going through, then it’s time to put it into the true production environment where it’s using real information. It’s doing real bookings. It’s answering real questions about insurance coverage, all those things that people are probably asking about that annoy the humans that have to answer those phone calls. We hear this thing that prompt engineering is the new literacy. What are your thoughts and maybe some best practices on how you get your team competently having productive conversations and effective conversations?

Yeah. Back to that point, prompt engineering, it is the new literacy. It’s how we talk to the machines and how we make them useful. What’s interesting is the more we stop talking about it as prompting because that sounds like an engineering term. That sounds like an architecture firm. The more we start talking about having conversations with this system as if it is a human or even if it’s treated as if it’s your number one employee, your number one coach, your number one consultant, your number one data analyst. You have a conversation about what success looks like. What are the goals? What’s the vision of the company? You’re teaching it more than you’re able to get more out of it. As the systems get more advanced, you’re not having to do as much of the structured prompt engineering that you may have had to do before because it becomes much more natural as the systems evolve where you’re just talking to it.

Interestingly enough, the voice AI assistant for my company, I use that when I go on dog walks. I put in the AirPods. I talk to my voice AI assistant when I’m taking my dog out for a walk or I’m driving from somewhere and I have a thought pop into my head because thoughts don’t always pop into my head when I’m in front of a computer being a keyboard warrior. But when I’m out and about and I have a little bit of fresh air and can think about things, I start having a conversation with my voice AI assistant. I don’t think about it as prompting. I think about it as having a conversation with this thing that’s trying to make me a better, faster, and stronger CEO. When I do this and I’m helping other leadership teams think about this as a conversation, they start to see a lot more positive impact coming out of those systems.

Off the top of your head, any advice to agencies to help them talk about and introduce AI to local business customers without panic inducing tendencies that technology can sometimes induce? The first place I like to start is asking people if they’ve played around with any of the AI tools and which one those are. Most of them will say, oh, yeah, I played around with ChatGPT or Gemini. You say, great, let’s do an exercise. Let’s imagine that you are the ideal prospect and you’re using ChatGPT or Perplexity or one of the tools instead of just doing a typical Google search. Let’s search for a problem that you can solve. If it’s a dentist, let’s use ChatGPT to search for the dentist in the area based on a person or their profile or what type of specialty there may be and see what those results show. What I found is business owners who are always looking for expanding their market share, getting more customers, competing better, if you just get them thinking about generative AI and the tools like ChatGPT as looking at it from their customers lenses versus their own lenses, that removes the science fair project aspect and turns it into a real business problem to solve. That’s one of the

areas I start and it’s a fun exercise to do. 2025 is an interesting year for generative AI and the adoption and 2026 is really going to be where we go into sixth gear. If you’ve got six gears in your car, we go full throttle.