The Agentic AI Revolution: How We're 10x-ing Our Productivity

Speaker A: This is Dave Learns AI, a tactico solutions podcast. I'm Dave Clapper, diving into the real world ways AI is changing business creativity and how we work. No fluff, no jargon, just practical insights and experiments you can learn from right alongside me. And here we are, back at it once again. You know, it's so interesting.

Speaker B: It.

Speaker A: It's so interesting how I can start a day with you guys and then end up when we do our podcast and being completely three different. Like, three, uh, different locations. Like, wait a minute. Weren't we just together a couple of hours ago?

Speaker C: Yeah. Uh, and now here we are, all

Speaker B: separate, put together again.

Speaker C: I got a new ring light. Check this out.

Speaker A: That's how you do.

Speaker B: Wow, it's nice.

Speaker A: I got one, too.

Speaker C: My only fans. Oh, nice, Dave. Nice.

Speaker A: I am broadcasting live right now from the switchyards downtown Atlanta. Uh, what about you, Nick? Where are you located?

Speaker C: I'm broadcasting live from Old Fourth Ward

Speaker B: and Tribble, and I'm coming at you live from the Sandy Springs Buckhead line.

Speaker A: Here we are. So the last couple of weeks, uh, we've had some guests. We've been fortunate enough to have some guests on, uh, Dave Learns AI. And last week was a lot of fun. A week before was, too. Um, it's always good to get perspective on other people that are trying to utilize some of the tools that we use in our daily now. And, uh, just the constant evolution of what is happening. It is. Nick and I were. All three of us were actually sitting in a meeting, and I showed Nick something that I was working on for a prospective client, and we started talking to this client about nine months ago. The client's a little tough to pin down, but he definitely is in need of some of the things that we can offer. And, you know, I was just thinking back nine months ago when we were first having the conversation going, yeah, man, we can do some AI stuff for you. And he's like, well, I don't really know what that means. And in all honesty, I didn't either.

Speaker C: Like, oh, we'll figure it out.

Speaker B: I don't know.

Speaker C: We're going to do some AI stuff.

Speaker A: We're going to make life different for you. And here I am showing Nick what I had done, um, through some of the tools that I use that not only utilize the tools that I've been mastering, but then bringing in some of the tools that Nick's been mastering and then combining them together into a presentation that now a prospective client can look at it and go, oh, my God. Uh, like, it has you know, they were talking about this revolution of AI and you know how digital, like the web, it slowly got there. You know, it was like, I remember my first WWW visit and then to where it became a daily thing, I feel like. And, uh, maybe it's just because we're the front lines, but this has gone from zero to a hundred like that,

Speaker C: zero to three hundred. So openclaw. Openclaw is one, one of our most downloaded episodes. And how do I know that's because we built an agent called Producer Paul that is now our podcasting. He's an AI podcast producer for us. So, um, he's gone. He's going through and auditing all of our most recent, uh, episode. Actually all of our episodes to see what, which ones drive the most volume and which ones are the most interesting for people. And openclaw is one of them. And openclaw, I have been neck deep in since it came out. And, um, I told these fellows yesterday it had updated seven times in ten days. That means that think about iOS, right? Uh, it's literally an operating system just like iOS. This updates once a year. Openclaw released another update last night. Eight times in 12 days. Eight, 12 days. And each update is significantly better. I've got hours trained because when you update, if you don't build it right, it breaks. So I've got ours trained to only update if the Update provides a 10% bump in productivity. And all of them have. Every single one, every one of them. So it's insane. Um, and that is how fast. So Claude code, all of these AI, they're all learning from each other. They are all learning from themselves. AI has now got to a point where itself improves. So I've asked our, uh, the Slingshot squad is what our agents are called. There are an army of AI agents, and they each are tasked with getting better every day, which is insane.

Speaker B: So it's like I like to think of things in paradigm shifts. And we talk about. You're talking about how far we've come in nine months since you talked to that prospective client. The most recent paradigm shift is the agents. Openclaw Manus has agents. Where it went from Chatgpt and all these LLMs used to tell you what to do and they sounded really smart. And they, you know, you could go do those things and be successful. This agentic paradigm shift has been. Now it does it for you and it suggests things to do and then it go does those and then they get better on themselves. So that, Dave, you were talking about how you were Doing nine things at once. You couldn't do that unless you had them working. You set it and forget it almost, right?

Speaker C: Yeah, yeah.

Speaker A: Or check back on it and go, okay, where are we at with this? You know, and then that, that's the thing. Like, uh, you know, I'll be going, I'll be doing one thing and then I'll go and get distracted with another thing and I'm like, oh yeah. And I'll go back and I'll go, oh, okay, cool. That was done. Awesome. Thank you. Now I can take that and go and do this thing now. So I don't think I go ahead

Speaker C: and because of that, we've also developed a task, uh, management platform completely built, uh, using Claude code. So basically ripped off Asana with Claude code. So we've got a task management, uh, platform for our AI agents because we have them doing so many different things. We're just humans, we can't be this attractive and smart people. So we have it all logged in, uh, to where the tasks are actually on a Kanban board, um, that our AI agents are working on. So if there's a human blocker, which API, like I have to give API access to some of our agents, they're waiting on me to give them the API keys for that stuff. So it's on my to do list. So they're like, nick, what's up? And I get pinged. We have a telegram set up where I get direct message from them. I get daily updates. My, my morning. I have cron jobs, which cron is short for chronological or chronograph, but it's basically time based jobs. So it's daily or weekly or monthly or however you set it. Task multiple per day. And I wake up in the morning and I go through my rundown of all the stuff they're working on, what they need from me. This, you know, that day. Um, what kind of. We had, we had a couple meetings today, three really big meetings and they prepared all of the dossiers for us. It's just, it's insane what's happening.

Speaker A: I wanted to pull this up real quick because this is one of the things that I was showing you. Um, you know, we were talking about AgentIQ and these agents that are being created that Nick is really working within the system, OpenClaw and creating, you know, professionals for all the different jobs that we need. So one of our clients is a retail location. They've got 24 retail locations around the Southeast, um, an outdoor brand. And what we did is we put together a team that is specific to their business. If you, if you are watching this on YouTube as opposed to listening it, uh, you may want to go over to the YouTube and kind of get a look at this. But you can see each one of these agents does something different. So you got a merchandise strategist, a local store marketing person for all one of you know, for each one of the stores, uh, Taurus acquisition specialist, influencer and ambassador programs, uh, reputation and reviews, digital advertising campaign calendar. Each one of these agents does a separate job. And that's where, I mean, that's just where things have gone, like, next level for me. It like, okay, I can bring in a client and go, all right, listen, I'm not your SEO guy, but I have one. I'm not your, I'm not your content creator, but I have one. You know, um, that kind of stuff that has just, you know, gone next level. And, you know, we were talking off camera a little bit about what a lot of this means for software in general. Um, if you think about, you know, being in business, I was talking to a friend yesterday and I was in radio with him, and we had software that we paid for, you know, a scheduling program, you know, something, uh, like that. And that software was only as good as it was. I mean, you couldn't really change it. Once you got it, you paid for it. And you're like, well, I use some of the tools in this. Um, but now you can create software based specifically on, on what your needs are without having.

Speaker C: They probably can't. They probably can't.

Speaker A: We can't.

Speaker C: We can't. They probably. It's. It's hard. It's not easy to figure out. I've spent three weeks, you know, in the, in the, in the mud with it, trying to figure it out. But, um, it is getting to the point, you know, HubSpot is hard. Salesforce is hard. You basically have to get an engineering degree to learn how to use them. And then even when you do, there's so many different features that you don't use, so you're spending all this money for three or four features that you actually use. What is going to happen in my prediction for the future is every small business and every medium and large business probably are going to have their own custom built by themselves or by someone they hire, like us to do it. Their own custom software that bridges the gap for everything. And it's one hub where all of their tools live, where it tracks their performance, their KPIs, all the stuff that they care about. Amazing for real estate agents, people that work independently and need a team keeping an eye on this stuff. This is where we call ours the Slingshots, because they are the weapons that the Davids use to bring down to Goliath. And we didn't even come up with that name. They came up with the name themselves because our why the tactical Y is to help small businesses compete, um, at the same level as those larger companies. So they heard ROI and they said, oh, we're. We're the Slingshots. We're what David used to take down Goliath. I'm like, that's really cool. Awesome.

Speaker B: Yeah. And Nick, can you speak to, uh, how they, like each agent has a specific role and what it's an expert at. Uh, can you speak to why they need to exist as separate entities as opposed to just going to one agent and having it do all these things like why, what's the separation of each person?

Speaker C: I'll ask a very easy question for anyone. Would you go to an eye doctor to get, um, get surgery on your foot? No, you would not. Same deal, right? It's. Would you go to a, uh, would you go to a foot, uh, doctor to get brain surgery or a heart doctor to your brain? No. So they are all very, very well versed in their perspective lanes. Now, one way that helps is because it gives you multiple perspectives. Because they all have their own thoughts, they're all, you know, they're all trained on different things. So as opposed to having one set of eyeballs on it, Electronic eyeballs, even. Even, um, you've got a bunch. And it. So the way that we have ours built, uh, we have a Discord server. It's internal use only, that we go and it's called War Room. And we'll say, hey, we need something so and so we need a dossier for this guy. We're meeting a new prospect. Or we need a proposal based on these meeting notes. And it. We have one named Bo. You guys have met Bo, but he's the Orchestrator, and he's literally the chief agent officer for Taptica. And he goes and he calls on bots or the agents that, uh. Yeah, they're not bots at all. They're agents that, uh, that are specialized to, to do the things that we would need to execute on that stuff. So again, it's like Ocean's Eleven, right? You have the guy that's good at the bomb stuff, you have the guy that's good at the planning, you have the, the negotiation guy, all that stuff. Um, that's why. So it's just a clear differentiator and path.

Speaker B: And the other thing is they can work independently. Like you can assign 44 different. Yeah. Simultaneously. You can assign 44 different tasks and have them all working instead of it both thinking through each specific task by himself.

Speaker C: Correct. And they're all. This is also a good way to um. So API calls cost money. So anytime that your agent has to go to a third party vendor or software like an Asana or mailchimp or a Claude, um, it costs money. So you can set the less strategic ones that don't need deep reasoning or deep thinking. You can set those if you're using Claude to use Sonnet4.6 or haiku. Right. So it's a lot less expensive. But we'll use BO with Opus 4.6 because Opus is the deep thinking model. Um, so all that stuff plays together and those models change drastically all the time. So we also have a way where uh, and this is one of the new updates that Overcloud just put out that it basically it directs the model so it will tell you which model would be the best suited for this specific task for this specific agent. So if we have them doing some deep, deep strategy stuff, they'll use Opus 4.6, um, which is the most expensive. Or if it's just like replying to us, messing around on Discord, it's pulling from Haiku, which is literally like that's a difference between spending $100 a day in API calls. And that's happened. We do that, um, we have done that to $9 in a day on API calls.

Speaker A: So you know, it was interesting. You were actually doing some work on the back end to find out what it has cost or what the cost would have been to put the portal together that we did for one of our clients. And average cost I think for what we've created is anywhere from 20 to $30,000 worth of work that we've done, um, in six weeks, uh, for this client. So it's, it's amazing what we can do. There are a couple things I want to cover. I want to cover, I want to get to the ChatGPT or the OpenAI versus Claude and the Pentagon just because, you know, that's something interesting I want to learn more about. But Tribble, you, uh, had a big week this week. You presented to a client. I'm curious to know a little bit about your experience of not only setting up and being prepared for that um, meeting, but also the tools that you used, um, pre meeting and now post meeting and how it's kind of affected your workflow and you know, what you've, what you've noticed and enjoyed about the process or are frustrated with.

Speaker B: Yes. Um, what are they talking, what do they say is the mother of invention? Necessity is the mother of invention. Well, in a short, short enough time period, we knew we had a meeting, you know, two days in advance. Uh, in advance. I was, I'm mostly in the chat gp. I'm mostly in the LLM world. You guys are operating more on that agentic side a lot more than I am. But having to really put together something that was formal and a presentation that we're proud of and obviously the level that we operate on, I had to kind of get in that, get in the mud with you guys. And so what I noticed with working with Beau through Discord and that side of things is it's just a level. It's kind of like what chat GPT is to Google search the agent is to what chat GPT is. It's just that extra level of research, the extra level of thought. Um, you're, you're just a little bit more wowed. And so me as the human in the loop kind of putting it, everything together, it, there's less iterations needed and they just, they're more advanced. It really knows more about what I'm looking for. And this goes back to prompting and I've been talking to AI tools now for 10 months or even longer than that. But so it's still important that you know how to prompt and how to, how to talk to them. But at the same time, both before the meeting setting up the uh, proposal and then after giving it the meeting notes and asking it to refine everything so that we can actually get a signed contract, I was just, you know, through using the agents mind blown. Paradigm shifted for me.

Speaker C: Yeah.

Speaker A: You know, one of the things that the prospective client was talking about was a tool that they use for lead gen and uh, it's a specific tool they pay for leads. And um, you know, there's a company that does it and I guess does it well. And what's interesting is almost immediately following that meeting we went into some of our tools and said, hey, could we recreate this in a way that could give the same kind of success that they're having utilizing the service without having to pay that service? Because we can build that from scratch based on, you know, I mean one, looking at what that company does and then two, just emulating it or recreating it in a way That m. Now this client, prospective client can have access to that without paying that company all that money. So like, I mean, and that's huge.

Speaker B: That's why Nick says short all SaaS software, ah, as a service company because it's so easy to literally cater the perfect piece of software that's easy to understand, easy to make, easy to understand, easy to fix. And it's really customized to exactly what their needs are. And you can slap a white label on it and call it yours.

Speaker C: And I think that, you know, it's a custom home. So you're building a customer workflow or custom tool that you're going to use based on how you're going to use it and what you need. So like, um, one of our, one of our clients, we are, we had this whole, I call it the super app for them, this whole dashboard and we had a working session within the day. They're super busy, they're all over the place. So our meetings with them are, get in here and try it out and let us know what you need, what's missing, all that stuff because we don't have the context or the expertise on their industry. And we did that. And I just sent an update that was this morning. I just sent a complete overhaul of their tool based on all the feedback. We record our call, we feed that into our AI, it pulls out the request. Then I work on building it, fine tuning it, going back and forth, making sure everything's good. And obviously we have skills in building tools and marketing and brands and all that stuff of our own, so we know what to ask and, and we could create it very quickly. Um, but it does not mean it's low quality. It's insane how, I mean we did, I don't know, it was, I saw the, the recap was like 139 little updates.

Speaker A: Wow.

Speaker C: Yeah. Based on. And in one day, you know, and it's, it's just crazy. It's stuff like, like, you know, tagging the project owner in this or being able to filter out through the project owner that's. Or, um, being able to see the overall portfolio of other projects and what is ahead, what's behind what's on time, budgets, timelines, phases, all these different things. And this is, it's, it's really crazy. But I know every company is going to have their own custom built software, um, in the next 18 months tops.

Speaker A: So our podcast is Dave Learns AI but I'd like to ask each one of you guys something specifically you learned this past week. Uh, with it when it comes to tools or uh, abilities or something that you're like, wow. Like Tribble, you got a wow moment from this past week.

Speaker B: Yeah, I think I just, just talked about it really kind of my. I mean I had some conversations with our main agent Beau, but I hadn't really needed him and, and see him, seen what he was capable of and that's what pre meeting and post meeting of a big, you know, client proposal meeting. Um, seeing his capabilities was kind of my wow moment. And then some of the things that you guys have been working on inside the portal and the software and the mission control. Nick, I've been seeing ads of dudes talking about what you did a week and a half ago. So like uh, putting all these things together, how they all come together. But I guess dance short answer to your question, Dave, is agentic AI is the 10x from regular LLMs. And I think people will continue to see that.

Speaker C: Yeah.

Speaker A: What about you, Nick?

Speaker C: Mine just the speed of improvement. The fact that openclaw is updated eight times in the last two weeks. Um, the fact that I get a lot of my. I watch YouTube all the time. Right. So I try to see what other people are working on and it's kind of. It actually was flattering because what Tribble just mentioned is these, you know, tech influencers on YouTube are doing stuff now that I did two weeks ago and what. Um, so I'm like, hey, on the right track.

Speaker B: Okay.

Speaker C: Look at me. Yeah, okay. But besides that, um, and it's weird. My neighbors are like, shut up. Because I do that all the time. But, um, but I would say like it's. I will literally look when the video was posted, if it was in February. It's obsolete. Yeah. It's in. It's. And um, I'm not exaggerating if it was, you know, if it. Especially if it's an Open Claw video, if it's from five versions ago, it uh, no longer is applicable. And I don't know how people are going to keep up with it, um, that have real jobs, that have their own jobs they're working on. Um, if only someone would start a company that learned AI. So they didn't, you know.

Speaker A: But we know those guys.

Speaker C: Yeah. So that like the two things that I think are going to be major vast improvements are one is the custom software builds and it can live locally. And I think that's something that we're going to be doing a lot of because like you're having your own dashboard. Right. Everyone always wants that. It does not matter what industry you're in, you want your own company dashboard for task management. Having all everything in one spot. Documentation, um, just everything. Comms, scheduling meetings. You can literally search all of your meetings to wait. When did the client say that he would pay the extra five grand? Oh, it was in this meeting. And then you could pull up the audio clip of them saying it. If they say, I never said that, you can literally send them the audio. Um, that is where we're at. But ah, aside from that, the slingshots that we're working with and it's. This is hot out of the oven, you know, still hot technology. So it does break. And a lot of my time is working on fine tuning things. It does forget things. For instance, completely missed the name of, of one of our clients this morning as I'm working, I was like, we're working. Where did you even come up with that name? So it still needs to have human. It's not perfect, but if you have someone that's paying attention to it and updating it on a daily basis, um, it's fricking unbelievable. And it's just since episode one, it's just going to get better. And I feel like we are at the Hockey Stick where The better is 10x better every month or even a hundredx better every month.

Speaker A: Let's, um. I want to talk about the news. You know, I'm, I typically put my head in the sand when it comes to a lot of current events and news stuff. I've just, I'm busy, I'm busy working. I've got kids. I got kids in college. I got little kids. It's just I don't have time to pay attention to the stuff that oftentimes just brings me down. But because I'm in this world now, I have obviously seen a lot that's been in the headlines in regards to Claude being, or anthropic being thrown out of the White House and Donald Trump going, all right, we're not gonna, you know, you're pretty much forbidden. And then chat GPT or open AI gone. Well, we'll come in. And now there's this whole mu move of quit. Uh, GPT. Like there are a bunch of uh, you know, anti, uh, anti Trumpers I guess or whatever who are like, okay, we're, we're, we're not going to use chat GPT anymore. Nick, you're more connected in tribble. Probably you're more connected with me as well, but in teaching me a little

Speaker C: bit like what's happening so anthropic they are the um, they, they're positioning themselves as the, the safe bet. So OpenAI was supposed to have been open source. That's why it's called OpenAI. Elon Co founded uh, it with Sam Altman and some others. Right. Elon left because they wanted to turn to for profit. Anything. Open source is not for profit. It's for, it's for the betterment of

Speaker B: human, the greater good.

Speaker C: Right, Right. So as soon as Sam Altman switched his tune to we want to be open source to know, we want to be closed API. It should be called Closed API now because now they're going after um, now they're going after these uh, the money. Right. They're trying to make a lot of money on it. It's totally fine. We're, you know, we're capitalists as well. Um, but anthropic, their whole thing is safety. Um, they recently had one of their, their high ranking like security people uh, quit because they were like, I, he's I, we're witnessing the end of the world. He's that worried about it. Um, I'm not that I'm not risk averse whatsoever. But um, Anthropic, the CEO they're all about and I understand they're like, we don't want our AI to be in control of autonomous drones that kill people. We just don't want our technology to be doing that. Which I completely understand. I would get that as well. Um, but OpenAI is like, we'll do it, you know, because they're chasing the dollar, they're about to IPO, you know, all that stuff. They just actually.

Speaker A: 700 billion.

Speaker C: Yeah.

Speaker A: Evaluation. 700 billion, yeah.

Speaker C: So they're not going anywhere. Anthropic will be okay too. The coders are mad at Anthropic because Anthropic keep doesn't allow you use OAuth. And OAuth is the cheapest way for you to build these tools that use the AP or not the API, but the models. Right, the LLMs. So, um, ChatGPT allows you to use OAuth for OpenClaw, which means that you don't have to worry about tokens as much. Every API call costs money outside of what you spend on your paid account. So we have Claude Max, we still have to pay for the API calls and that's the stuff that racks up. Um, in order to skip the API calls you can use OAuth, which is basically just giving it carte blanche. Um, OpenAI says you guys can use RO off. So coders like that because It's a lot cheaper for people that are building things right now. Anthropic wants to make that, that money. So they are putting up the third or the wall where it's not allowing coders to do that. So it depends on who you ask. And I say, hey, they'll figure it out. They're all going to find their lane right now. It's the wild West. Is it? If, if we go in and we say, hey, uh, Claude code, we want to rebuild N8N but have a custom version of it for ourselves. Is that copyright infringement? Is that a, you know, is that something, uh, that we could be sued for because we don't have a patent on their software? I don't know and I don't think anyone knows yet.

Speaker A: Yeah.

Speaker B: Ah. And I think it's just a, uh, I'll just say it's a sign of the times. Like when you have the most powerful person in the world denouncing something, it's time to choose a side. Or you have to choose whether you want to support. Because if you give access or if you give the people, uh, empower what they want, where does it end? Like, you know, Apple. Apple always talks about their privacy, but at some point they let, they let FBI start getting into their unlocking phones for their kind of purposes as well. So it's kind of like you either stay kind of the bastion of freedom and of safety or you eventually kind of cave. And it looks like that anthropic's trying to carry that banner.

Speaker A: Yeah. Well, it's interesting. We'll see how it all kind of shakes out. Um, you know, it's kind of funny. We jumped on this call, kind of, hey, what are we going to talk about? And then we just jumped in and talked for about 30 minutes. That just goes to show you, um, not only, not only that we're really into this, but we're also really enjoying it. I found myself at 11 o' clock last night just in the middle of building something and just being excited, not wanting to put it down. Like, I got my music on and I'm playing on the computer. TV's not on. I mean, some might say it's not healthy, but it isn't. A lot of it is a lot of fun.

Speaker B: You're old cipher Puck.

Speaker C: Yeah, absolutely. It's just fun to build. And some people are like that, some people aren't. But I love to build things. And once things are built, I move on typically. So what's cool about what we're building here is a company that builds things for other companies. Um, so we're always going to be building, obviously getting ourselves better and growing our business, but growing other people's businesses as well. And I'm pretty much any corporate. And so we are trying to arm all of our clients and customers with the tools that will allow them to compete at the same level, if not better and more agile than the Fortune 500 companies. And I just think that's awesome. Yeah.

Speaker A: Well, uh, good talking to you guys, and, uh, we'll probably chat tomorrow. Another edition of Dave Learns AI in the books, and producer Paul will help us package this up and send it out to the masses and, uh, tell us what we did wrong and maybe hopefully make it even better next time around, I guess. But thanks, guys.

Speaker C: Thank you, Dave.

Speaker B: Thanks, guys.

Speaker A: This is Dave Learns AI, a tactical solutions podcast. I'm Dave Clapper, diving into the real world world ways AI is changing business creativity and how we work. No fluff, no jargon, just practical insights and experiments you can learn from right alongside me.