Bot Squads: What Happens When Your AI Tools Finally Talk to Each Other (with Kelly Sinclair)

Monica Froese [00:00:01]:
Hey, welcome back to the Empowered Business Podcast. I’m Monica Froze, and today’s episode is a little different. I recorded this live at AI that Sells, an event I hosted where I bring in people who are building things with AI right now. This session is with Kelly Sinclair. She just launched a new platform called Wave and she built it to solve something a lot of us run into once we’ve been building with AI for a while. If you’ve made more than one AI tool for your business, you already know they usually don’t talk to each other. You end up sending people from one to the next, hoping they follow the map you drew for them. And if you ever wanted to sell one of these tools as its own product, there hasn’t been a great way to do that without just handing someone the whole thing, instructions and all.

Monica Froese [00:00:45]:
Kelly walks me through exactly how she’s solving both of those problems. She gives us a full demo of the platform, and then I get to ask her all the questions I know you’ve been asking too. Honestly, some of what she shows me changed how I’m thinking about my own setup and I think it will do the same for you. So let’s get into it.

Kelly Sinclair [00:01:04]:
Oh my gosh. Good morning, Monica. Thank you for hosting. So I’d like to introduce myself. My name is Kelly Sinclair and I’m the co founder of a brand new AI platform called Wave and by Gravia Studio, which is our, our company’s name. I’m 100% accidental tech co founder. I. If you asked me a year ago, like Kelly, in April 2025, I was not on this journey.

Kelly Sinclair [00:01:32]:
I was like, oh yeah, I’m playing with AI. It’s fun, it’s cool. I got an Instagram memory that popped up the other day and it was February 2025. I had just built my first custom GPT. So I’m telling you this because you can learn a lot really fast when you start focusing on AI thing. And my trajectory was I knew this was going to be important. I knew that I wanted to start adding AI into the way that I provided my service as formerly a marketing and brand strategist. So I was working with people like you, entrepreneurs, course creators, experts and whatnot, to help you, like, figure out your branding, figure out your visibility plan.

Kelly Sinclair [00:02:17]:
And what I was finding was my clients were like, cool, nice strategy. And then it would be like, but, but how do I do it? Like, how do I implement that? And that’s where I saw AI really coming into play, was the opportunity to really support with the Implementation, making things easier. Like, you want to get visibility, you want to get on podcasts. Well, you need a bio, you need to know what podcasts, you need to understand how to pitch. And we can build AI tools that will do all of those things and will help you. Right? So that’s where I was really seeing it coming together. And. And then I started having some coffee chats with a local software developer and sharing all of my exploration and nerding out about AI things.

Kelly Sinclair [00:02:58]:
And he had been in generative AI for the whole time, like three years. Like since ChatGPT became a thing at the end of 2023, it was. And. And we eventually were just like, well, let’s. Let’s fix this problem that I had. And I’m going to talk about all the things around custom GPTs, not talking to each other, not having a really secure way to actually sell those custom GPTs that you feel good about as a creator and whatnot. And. And then we decided to just start a company.

Kelly Sinclair [00:03:28]:
So no big deal. I’m also a Canadian. I live just outside of Calgary, Alberta, and I side hustle as a professional mom. Taxi. So after this, I will be shepherding my children around to gymnastics, and we’re about to get into softball season, where I will be six out of seven days of the week at a softball diamond somewhere. And hopefully it’s not snowing still. All right, so this is what we’re talking about today. And if you are any of these things, you’re definitely in the right place, or if you’re thinking about anything, these things are definitely in the right place.

Kelly Sinclair [00:04:01]:
So you’ve already built like multiple custom GPTs inside of your offers, your memberships or your programs or. And you’re selling them. Right? You have paying clients who are actively using these things, and you are experiencing delivery friction. Delivery friction, meaning how do I actually share access to this? How do I turn it on, turn it off? How do I, like, securely feel comfortable baking my ip, all my frameworks, everything that I’ve learned into an AI tool and then letting somebody use that. And you really want to scale this, right? Because this is like the way. The way of the world right now. And I know Monica has like a hundred billion of them. She’s like, on the leading edge of this.

Kelly Sinclair [00:04:40]:
So it’s happening, folks. Here’s what most creators are doing right now. And I would love to see in the chat if. If any of these are you. So you’re building a custom GPT, you share the link. You don’t even think twice. About the fact that people have permanent access to your tool and they could probably just share that with other people as well. Option B, you’re embedding a password.

Kelly Sinclair [00:04:59]:
You’re like, okay, I’m going to add a layer of security to this. I’m going to embed a password. I have to manually update it now I have to remember to tell my students what the new password is. And that creates its own project management challenges for you. Option C, you’re like, I’m building a separate GPT for every single one of my like individual one on one clients and I’m going to delete that GPT access if they cancel or they graduate or whatever that looks like like, but then you also need a like tracking system for that. And option D, you’re, you’re seeking out some alternative solutions and these are higher cost things to add on to what you’re creating right now. So maybe like a rad genius or like a pickaxe. Okay, so, and then also I have to address this because Monica just spent a whole hour talking about Claude skills and everyone’s like mass exiting chat GPT and going, I want to do skills.

Kelly Sinclair [00:05:50]:
How do I share my skills? You’re like, well, if you were to share your skill the same way you share a custom GPT, you’re literally handing over a zip file with your frameworks and all of your goodness baked right into it. So you question yourself, do I really want to do that? This is, this is an IP problem and it’s the elephant in the room. And that’s what we’re going to talk about. Because when you hand over your expertise to AI, the way you deliver it does matter. And I think that for a while, like when a solution doesn’t really exist, we kind of just accept, accept that this is the, this is the risk that we’re taking and was probably affecting the way that we might price things. How we’ve decided what we’re going to put into that, how we’re even combining using AI tools and like live touch points or other supplementary materials like blending them into a course or whatnot so that you can kind of like have AI just do little bits. But like the, the Big Kahuna Stars still falls on you. Custom GPTs love them.

Kelly Sinclair [00:06:50]:
And also we love them. And right. So on the left you’ll see this is like the reality of custom GPTs. Right now we’ve got your IP floating around and shareable links. You maybe are doing this manual password nonsense. You’re locked in to OpenAI and their models and also their political choices, which is the. That’s a whole other conversation. You also have this frustration around.

Kelly Sinclair [00:07:17]:
They don’t talk to each other. Right. That’s something that we. Monica was just saying about how the quad skills, like, are system wide and they can always be accessed and they understand, you know, they know when to get pulled in. And I can tell you, like, I was building this is one of the reasons why we created Wave was I was like building a flow for my podcast, but I’d have to go into one custom GPT to take the transcript and have that turn it into show notes and then another one would turn it into an email and then another one would turn it into a social media post. And I was like, it’s so annoying that they don’t just like all work together. So that’s something we were trying to solve. And then also.

Kelly Sinclair [00:07:56]:
Right. You’re like pushing someone to ChatGPT. ChatGPT has never paid you for that either, by the way. Right? How annoying. How rude that they don’t have affiliate links for all the people that you’re sending to use ChatGPT as a result of. Of your tools being available as custom GPTs. So I kind of like wanted to put a wish list here. Here’s what I was wishing for and maybe how you’re feeling as well.

Kelly Sinclair [00:08:18]:
You want gated, controlled access subscriptions that you can turn off automatically. You want the ability to use any LLM and switch when you want and not just from a, like, how do I align with the organization behind it? That’s definitely something. But LLMs work differently and they all have different skill sets, different like things that they’re best at and you might want to use different ones. Also costs, right? Like token usage is a thing. So being able to be flexible on that, it’s really helpful. And you want connected bots that have shared context. You don’t want to have to explain things again and again and waste your time and feel like you’re like working with an intern that’s not paying attention to you. And also, what about the ability to brand it and, and have an experience that actually feels like you and that it’s acknowledging that you created this thing.

Kelly Sinclair [00:09:09]:
So I have to take a moment and address this market opportunity that we’re in. Like we’re seeing where this is going. It’s very familiar. It’s very much, I’m calling it the like, online course explosion circa 2026. So it’s like it’s happening again. Right? Remember 2020 and 2019 even when this was kicking off, lots of you were probably there. Maybe you were either building online businesses or you were buying up all of the online courses. But that market went from 45 billion to 165 billion over the course of two years.

Kelly Sinclair [00:09:44]:
And the early movers captured 70% of that. So if you’re on the front edge of that, you made bank and kudos to you. But then there’s this issue where we saw this drop off, right? All of this information is being consumed or collected anyways, but not actually consumed consumed because we had a 10 to 20% course completion rate. So hands up if you’re like me and you bought something, but it’s like sitting in your digital course graveyard now, right? So we’re now moving into what I call expert backed AI as a product. So while courses teach people how to think expert backed AI tools, let them use your thinking. So this is the big step change between how do I help somebody do something by trying to teach them all of the steps and actually building something that does some of those steps for them. Some research has been done as well about what changes when you start mixing in AI tools. You get a 45 to 70% completion rate when your tools are guided by AI.

Kelly Sinclair [00:10:44]:
So that’s like three to five times higher than just like a typical digital course. And here’s the big one, it’s actually about time to result. So people, we live in this instant gratification world, right? People want results fast. So we used to think like a digital course that you walk through people for 30 days or, or for six weeks maybe was kind of quick. Now it’s like two to five days. They want results and I would even like venture to say two to five hours in some cases. Like sometimes you can get, you can purchase an AI tool and you can fully revamp something and have a result like almost instantaneously. Which is amazing because this is really about thinking about your clients and how you can improve the experience that they have.

Kelly Sinclair [00:11:32]:
So are you already doing this? Are you already building expert backed AI’s bots that guide clients through your frameworks? Clients. Then when we’re talking to our, in market research, we’re talking to people who are starting to implement this, especially in like a hybrid setup where they’re, they’re creating AI tools that are helping their clients and then their clients are coming into like office hours and coaching calls and they’re talking about how these conversations are just so much more nuanced and deep. So it really like allows you as the expert to Use your human brain and really showcase that even more. Which is something by the way, that I think is so important that we’re not just going like, let’s build everything AI, handing everything over to AI, let the robots do it. This is like, and this is a core value of mine and Andrew, my business partner at Wave, is that we’re creating infrastructure that supports human centered AI adoption. So what does this mean for you as a creator? Right, that you’re able to have these conversations do your thing, but also add extra value by having AI tools supporting your workflows and your clients then are implementing faster. Right, because they’re not stuck on the basics, AI’s handling. So that’s why we built Wave.

Kelly Sinclair [00:12:56]:
So this is a platform where digital creators can securely build and sell expert backed AI tools on any LLM. Let me tell you a little bit about it. And don’t worry, we’re definitely going to get into the demo, we’re going to see how it works, I’m going to show you some things and very excited. So our core kind of foundations are that it’s protected, you have, you can share the tools that you build with secure access. It requires logins, we have all the security features built in. Actually that’s just a, it’s learning so much joining, building a tech platform. I tell you, connectedness is all about the nerdy way to say this is multi bot orchestration, which means things can talk to each other and they know how, they know when they are supposed to do their job. And then the flexibility of having a variety of LLMs.

Kelly Sinclair [00:13:51]:
We are always staying on top of the different LLMs when they drop. Like we launched, we literally launched Wave one week ago. Okay, so I was finally able to say that name out loud one week ago. So if you’re here today, you are getting a very, very, very, very, very, very, very, very early start sneak peek of it. And of course Claude dropped Opus 4.7 the day before and we’re like, well, you know what, it’s still going out there, we’re still doing this and that’s what’s going to happen. We’re very much committed to ensuring that we are always using the most premier versions of models, like allowing access to that and also making sure that any models that we have on our platform meet our security and privacy standards as well. At the moment we actually are not providing access to Google because Google’s privacy policy, which may or may not have been updated, I will say that as a disclaimer, was saying that they would use your Data to train their models and we’re not about that. So no Google for us right now.

Kelly Sinclair [00:14:54]:
So bot squads. I was actually at your opening session, Monica, and you’re like, I don’t know what Kelly’s doing, but it’s something about bot squads. Because she keeps talking about bot squads. And yes, I’m trying to make bot squads a thing. Okay. Bot squads are going to be like, fetch. We’re going to make it. We’re going to make it a thing.

Monica Froese [00:15:09]:
Yeah. I was like, I felt like every time I tried to explain what it, what you were doing, I was going to botch it. So I’m like, we’re just going to wait for her to tell us. Yeah, hopefully I explain it.

Kelly Sinclair [00:15:19]:
So bot squads is essentially what the cute way that I like to call what multi agent orchestration is. And here’s the slide. This is like if the one thing that you can take away from this in terms of how our platform operates, it’s like custom GPTs and quad skills and workflows all had a baby. Just going to let that soak in. I was like, claude skills are awesome. And also that’s what we already built and why, like one of the main frustrations that probably a lot of us feel about. Thank you for the brain explosion emojis. That’s probably how a lot of us feel about the way that custom GPTs work.

Kelly Sinclair [00:15:58]:
We’re like, we love them, right? Here’s the things we love about custom GPTs. We can train them to do a defined job. Their output quality can be really good. You can give it the knowledge files and that that stays hidden and protected and everything like that. And it does its job really well. But it’s isolated, it doesn’t know what other bots are doing. It doesn’t have any context passing or a system. And then we’ve already talked about the IP sharing part of it for Claude skills, they operate at the system level.

Kelly Sinclair [00:16:27]:
They’re not locked into a single chat window and that context will flow across conversations. So it’s a capability thing and it’s a memory memory thing. But then we have the. We don’t love the idea of sharing skills necessarily, depending on how you build them, obviously. But if you think about, can I just convert my custom GPT into a cloud skill and then share that, that’s where I would say, I don’t know if you want to do that. Right. Like what’s in your files. And then we have workflows.

Kelly Sinclair [00:16:53]:
Love a workflow. Love when we can automate things like, and I’m going to show you this live in action with our AI tool launch playbook. But I have taken some, I know some people who are super into workflows and they’re super nerdy about it and they’re freaking cool. But then you’re starting to get into the territory of having to have all of these other subscriptions and all of these other costs, learning another new tool and JSON, like, let’s not even go there, right? Like, we need to learn a whole new tech language sometimes to effectively use a workflow and, you know, not, you know, some of us just don’t want to do that. So again, this is really about your client experience and what happens with a squad versus without one. So right now, and I know I’m going to use you as an example, Monica, because I know you have a program that has like eight or nine custom GPTs, but you also have like a map that you have to give your clients that are like, gotta use this one, then you gotta use this one, and then you take that and you put it in here, and then it goes here and then it does that and whatever, right?

Monica Froese [00:17:57]:
Like, I call it chain linking. But yeah, it’s not seamless. You. And that’s why I had to create a map for it. So people knew which one to use and when.

Kelly Sinclair [00:18:05]:
Yeah, yeah, exactly. So, so you’re having to like, be like, here’s the tool, here’s how to use the tool. Reference this document. Like, so it’s just kind of like not the most ideal experience for your clients. It’s just. It is what, that’s what exists right now. So like, well, let’s fix that, right? Yeah, let’s fix it.

Monica Froese [00:18:21]:
Yeah, exactly.

Kelly Sinclair [00:18:22]:
Let’s fix it. I want to fix it. So with the bot squad, you’ve got the one place to go. Context will flow throughout the system. It has a clear path, less repetition, more momentum. And that allows your clients to implement your framework more seamlessly and not have to be like, I can’t figure out how to use these tools. So we have a couple different ways that you can deploy this to. And this is just getting a little extra layer of nerdiness, one that we call workflow, which is where the bots will actually hand off.

Kelly Sinclair [00:18:51]:
And that’s what I’m going to demonstrate for you. The bot’s like, I’ve done my job, it’s the next one’s job, it will do its job, et cetera. And then there might be circumstances, of course, where you want like an advisory board or like, an advisor group where you might want to pick and choose which ones you want to use that you can manually tag and say, like, say you want to have a conversation with, like a marketing team, right? You want one that will build your marketing strategy, but then you might want other ones in the same room that can then know your marketing strategy and execute on developing those assets for you. And that can all happen. And I call it like being like in a group chat, right? Essentially, in order to build a bot squad, which is very meta, what I’m about to show you how to do using a bot squad that we built, because that was the whole purpose of the AI tool launch playbook, but it needs to think through these kind of steps, the first one being the input. So what information does it need to know before the first one can do its thing? Or is the first one collecting that information from the user? Then what is the process? What are the steps that we’re going to break this down into to actually get to a result, and what do we want that output to look like? So what, where do they end up? Like, what is the system? And I tell you, like, before AI, I never could really think very well in frameworks. That was not my. My vibe.

Kelly Sinclair [00:20:13]:
I was like, I don’t have a framework. I’m a strategist. I make strategies and they’re all different for every person. And then you talk to AI and it’s like, here’s the patterns of exactly how you do this. So AI is very good at identifying patterns and turning that into tangible, repeatable things. And that’s exactly what you need when you’re using generative AI to support a workflow, a process that you’re doing yourself or something that you want your clients to work through. So I’m going to walk you through how the AI tool launch playbook works. You’re all getting access to the AI tool launch playbook.

Kelly Sinclair [00:20:52]:
So this is like, first I want to say to you, this is not the full platform. This is just a tool that we built with our platform that tells you how to build tools on our platform. Again, super meta, very proud. And so this is what you’re going to go download, and then I’m going to demo this in a second. So it starts with, I tried to keep these names real simple. Normally I would be super cutesy about the names, but we have an intake bot that’s going to guide you through the discovery, is going to ask you questions, going to understand your program so that it can identify opportunities, which is the next one’s job. So the Clarity Bot is going to take. Okay, here’s what you do.

Kelly Sinclair [00:21:31]:
I’ve asked you some questions about where your clients get stuck, how we can think about where we can create AI tools, bot squads that might help your clients to go through those steps. So the Clarity bot’s going to distill it out into a little snapshot, and you’re going to go, yep, that sounds good. And it’s going to give you some ideas about where you might plug this into your current offer. Then we have an architect bot. An architect bot designs the full bot roster. She’s like, here’s everything that we need. And there’s like, some nuance. Obviously, the one thing that’s different about creating a bot squad to just creating a whole bunch of individual custom GPTs is that we have, like, another layer of instruction and guidance at the squad level.

Kelly Sinclair [00:22:18]:
So we need to be able to actually communicate what kind of handoff might need to happen and how these things all talk to each other. So not just the individual’s jobs, but how they live as a team together, essentially. Then we have the builder bot. Builder bot’s going to knock your socks off, because Builder bot, she is going to take that architecture and actually spit out the instructions and suggested knowledge files that go with every single one of these bots. So even if you don’t end up using Wave, you’re going to get a series of custom GPTs with instructions and that you can go and build in ChatGPT if you want to. You’re not going to want to, but you can. I wanted to make this useful regardless. And then I have a preview bot, and this one’s job is literally to just show you a little tour behind the scenes of how it looks as a creator using wAIV.

Kelly Sinclair [00:23:12]:
So we are going to jump in to the demo. All right, so you’re over here. So, Monica.

Monica Froese [00:23:20]:
Yes.

Kelly Sinclair [00:23:21]:
I’ve been practicing going through this using the. What you have described over here in the Empowered Business Lab on your empowering job.

Monica Froese [00:23:38]:
I don’t know. Does that actually even reference all of the GPTs? Because.

Kelly Sinclair [00:23:42]:
No, it doesn’t. But that’s a good thing because I know you have more than one.

Monica Froese [00:23:45]:
Yeah, that’s like, I haven’t updated that shop listing forever. There’s, like, a main sales page for it. Okay.

Kelly Sinclair [00:23:51]:
Yeah, so I was just trying. I was trying to find that, but just because I’ve already, like, got this kind of copied and saved.

Monica Froese [00:23:57]:
Okay, I’m very curious about to happen.

Kelly Sinclair [00:23:59]:
Okay, so I’m just in a chat window it looks very similar to cost, like ChatGPT or Claude. Right. And I can see all these bots and these bots are all built on different LLMs. So I’ve got Sonnet 4.645.4.5, Opus 4.5 and then this last one. She’s just on the haiku because that’s fast and she doesn’t need to think, she just has like one thing she needs to say. So right now you have to initiate. Make sure I’m talking the right one. I’m going to say hi, going to move over.

Kelly Sinclair [00:24:34]:
So this is what you’re going to experience when you go through the AI tool launch playbook yourself. You’re going to get this little thing that explains what a bot squad is. We just went over that. We’re talking about how to figure out what yours looks like. So to get started, do you have a sales page, a program description, something I can work from? Or we’ll just maybe you’re coming up with an idea, so. Or we’ll just like interview you. Here is my sales page and I’m copying your page that I just showed. So just for.

Kelly Sinclair [00:25:06]:
It’s rich. Monica, just so you know, credit to you, a lot I can pull from.

Monica Froese [00:25:12]:
So I have an.

Kelly Sinclair [00:25:13]:
I want you here because I want you to like validate the outputs just a little bit. But again, this isn’t about like, how good is this? You can add, you, I think you can have lots of conversation here. You can add files to this for extra context. It can really get juicy as far as like getting it to, to show you your opportunities. But I just mostly want to show you the, the process and the workflow.

Monica Froese [00:25:36]:
Yeah, I’m reading it right now. It’s writing copy. Yeah, it’s got it for. For what? For what that listing says it does. It’s correct.

Kelly Sinclair [00:25:46]:
Right. So what I’m wondering is whether there’s a specific moment in the student journey where things slow down, require the most from you personally. A place where, you know, a more connected guided sequence could do some of the heavy lifting. So where do your students get stuck or stall once they’re implementing this? That’s what we’re going to answer.

Monica Froese [00:26:07]:
There’s a lot of places people get stuck, hence why I have nine custom GPT. Well, you know what would be. Let me read to you the other six that I added into the program. So these are the core. So the program’s built on three pillars. And so the product creator is for pillar one, which is digital products at scale, funnels at scale. So it matches that. But then as I was getting in and creating these bots, I realized oh, I can plug a lot more holes.

Monica Froese [00:26:38]:
So let me tell you. Okay, so actually my process now kicks off with a reverse idea finder GPT because that’s actually what validates the product. Then they move over to the product creator. We also have an individual opt in GPT because your opt in logic is different than the logic of the actual product. We have an order bump GPT because that’s also different logic than the product or opt in. Then we have the funnel builder price like a pro GPT. People get in their head so bad about pricing.

Kelly Sinclair [00:27:12]:
Yeah.

Monica Froese [00:27:12]:
And I’ve like here I like developed my own framework around it because so many people get stuck. I’m like, it’s. Anyways, that’s the whole thing. Data analyzer is by far my favorite GPT because it’s like my data brain on.

Kelly Sinclair [00:27:25]:
Yeah.

Monica Froese [00:27:26]:
And that’s probably the biggest in terms of like people getting stuck. People hate data. And so I’ve always been good even before AI. I can get people to create their product, validate their product, do the sales pages and then they get to the data and they’re like, they lose their minds. Yes.

Kelly Sinclair [00:27:45]:
Okay. So I anticipated this answer. I was like, I bet it’s like well one place a definite opportunity anyway. And the whole purpose of this is to just find something like where is a stuck point? We’re not talking about like building an entire program necessarily with this, but it’s like how do I add to what I already have? And I was like, I bet there’s like they have a hard time testing and iterating their product once the first version of the law is launched. They need clarity about how to interpret and apply the insights that they learn once you know, they start.

Monica Froese [00:28:14]:
Yes.

Kelly Sinclair [00:28:15]:
So there we go. That makes so much sense. Totally is.

Monica Froese [00:28:20]:
It’s so not getting the data. No, they. They can do that. And pulled out knows I have a funnel tracking spreadsheet. That’s the easy part. It’s like and. And the. For all these years the tracking spreadsheet even calculated all their conversion rates.

Monica Froese [00:28:32]:
But still. And I could give them a million if then, then if this then that if your data says this then do that I could. But they still lost in their head. They like literally want someone me who can’t do this for a thousand students to basically like tell them exactly what to do next with their funnel. And that’s what the data analyzer GPT does.

Kelly Sinclair [00:28:54]:
This doesn’t exist anywhere else. Like the ability to program in that automation where the, where you are doing the programming. So what is suggesting? It’s like, here’s your opportunity. We’re going to make the funnel optimization strategist squad. It’s for your empowered business lab students who’ve launched their product funnel, had these things happen, but they’re staring at funnel metrics and can’t translate that into confident prioritize action. So it’s just identified a problem that your students may be facing. It kind of breaks down this process that I showed in one of the previous slides where we need to understand what the input is. So like, what are they bringing to the table? What do we need to do?

Monica Froese [00:29:37]:
This is exactly what it does. The one in my bill. They give it the spreadsheet. They give it. Yeah, this is accurate. Yeah, accurate. Good.

Kelly Sinclair [00:29:45]:
Okay, good. And then, and again, like I built this tool just to show you how to build this, these tools. It’s not really wAIV. Wave isn’t this tool. This is just a demonstration of how we could build a tool here. And I wanted it to be a tool that would help you figure out yours. And then so here’s what it’s recommending a funnel or. Yeah, so we’re going to have a funnel health snapshot prioritization recommendation, a testing roadmap, and it also needs to have a what not to do section.

Kelly Sinclair [00:30:14]:
It also says we could do a pre launch funnel blueprint bot squad or we could do a traffic strategy matcher. So it’s just like a couple of ideas. I’ll just say I like this idea. Let’s do it. Okay, so now I’m going to make you all watch when I hit enter here, what happens. I love your energy. Now, this is her. Please behave.

Kelly Sinclair [00:30:38]:
Yes. ArchitectBot has now been brought into the chat. She’s like, okay, here’s what we need to do. I’m passing this information over to the Architect bot. The Architect bot is going to tell you we’re making the funnel optimization strategist. These are the stages. This is what needs to happen. And now it’s popping out this bot roster, which is actually kind of defining the different bot roles that it’s going to create.

Kelly Sinclair [00:31:07]:
So I got three. While that’s building, actually, it’s going to have to build again. Okay, you already did that. So thank you very much for your assistance. So what do we have here? We have squad house rules. So this is what’s different because we need to give the squad some rules. Right? Be strategic, ask, don’t guess, lead with highest leverage. Action, Keep it actionable.

Kelly Sinclair [00:31:31]:
Respect the data threshold.

Monica Froese [00:31:33]:
Is that thinkophantic?

Kelly Sinclair [00:31:36]:
Is it a lot? I don’t know. We should look it up. You’ve never heard that word before, clearly. I maybe need to give some like, vocabulary restrictions to match the brand voice. So, like, you can give your squad your brand voice if you want it to be really like, that’s what you’re kind of thinking about sometimes we’re actually like kind of cloning ourselves and creating these little tools that have our brains in them. We want them to act like us a little bit too, as much as we can do. But I will say, like, if you haven’t heard the term non deterministic before, that’s what AI is. It means instead, even though you’re programming it, you have two variables at all times, the input and the output.

Kelly Sinclair [00:32:17]:
They’re always going to be different. So what, like, you might go through this. I’ve gone through this squad with your stuff probably 20 times for this demo and it’s created a slightly different version. Even if I say the exact same responses, you know, that’s just how generative AI works. So it’s like the grain of salt that we have to always consider when we’re building stuff. So it’s built out the funnel health analyzer. Again, we. You don’t need to read these, but essentially it’s like, this is what you do.

Kelly Sinclair [00:32:48]:
Here’s like a script or whatever, here’s some funnel benchmarks. Like, here’s what you might want to upload as knowledge files for this. Then we have bought two, the priority recommender. Here’s what you’re supposed to do. Knowledge files, any frameworks, like, it’s kind of suggesting if you were like, I don’t know what I should build, it would be like, do you have something like this or that? Like, this is what would be supportive for this bot to do its job well. And then the testing roadmap builder. And it has instructions again. So.

Kelly Sinclair [00:33:18]:
Oh, and it has a fourth one, a strategy document compiler that will do what it does. So there we go. And now we’re just going to go. I’m going to show you the next one because it got a little. So yay, the preview bot. This one’s job is literally just to repeat itself. So what I’m going to show you is a quick like 4 minute view of what happens behind the scenes. So the preview bot is like, okay, you built a squad.

Kelly Sinclair [00:33:48]:
Here’s what it looks like to set that up. Right? Because what I’m giving you in The AI tool launch playbook is just the ability to use this tool and it’s very much the subscriber experience. So it’s your client’s experience. When you have an account, as the person who built the Bot Squad, you have a creator account. So this is what I’m going to show you is what it looks like inside of a creator account. So tell me if you can hear this. I want to show you a little behind the scenes of how you set up and share a Bot Squad AI tool on our Wave platform. And then I want to show you what happens when your client follows the link.

Kelly Sinclair [00:34:34]:
So when your client clicks the link from however you have shared it, whether that was through an email confirmation after they purchase or directly embedding it into your course platform, they’re going to get this screen that shows them what they’re subscribing to the name of your tool, who it was by, and also a very clear indication that they’re on a platform called Wave, for which they now have to sign up. There is no cost for them to sign up at this point, we are using, allowing them to just use it for free, and we’re managing how that works out on the back end as part of our testing right now. So I actually already have an account, so I’m going to click sign in and go through my login process here. And you can see that it just launched me straight into what feels very much like any kind of AI chat platform. But it has your Bot Squad directly embedded in here, and we can start having the conversation with the first one now. So this, this part goes back and forth around all the different bots, and this is where it is built out as a sequential workflow. So Wave is currently in beta testing right now, so we are making a lot of changes and adjustments and optimizations based on feedback from our beta users, creators who are building and selling squads on our platform. And if you’re interested in being one of our beta users, then you can fill out an application by sending me an email to kellyraviastudio.com so we did that.

Kelly Sinclair [00:36:08]:
And now one last question that I’m anticipating before we can get into any Q and A is like, how do you move over here? Well, we built it. I hope you saw that from the behind the scenes to feel very much like custom GPTs. So what’s really cool is you could actually transfer any custom GPTs that you have and any Claude skills that you have, because Claude skills are still essentially the same kind of level of instruction you’ve got a Claude skill file and then you’ve got any reference files so they, they replicate just nicely over there as well. And what we’re doing right now is, is founding creators. So anyone wants to talk about that, feel free to connect with me. Here’s all of our connection information. The platform is wAIV-AI.com. we just launched a brand new Instagram with like seven followers.

Kelly Sinclair [00:36:59]:
So please come, come along and follow us over there and, or you can reach out to me directly. I also have a podcast called Entrepreneur School in the AI Era where we talk about AI in business and how we do it now. And it is, it’s a wild ride. Friends. Excited to be part of this with you. Okay, let’s do it.

Monica Froese [00:37:20]:
So I people honestly like they, they were like, I think very invested in watching the demo because like they’re like awesome, amazing. I have questions though, so I’ll go

Kelly Sinclair [00:37:34]:
because I feel like, oh, I know you will, I know you will. And everyone’s gonna benefit from them. So let’s do it.

Monica Froese [00:37:39]:
All right, so what you demoed for me as someone who’s created like a hundred plus custom GPTs, right? And I’ve created ones that should work together. It’s a little janky to do it in ChatGPT and I’ve created standalone ones and now I’m as you saw, I’m like getting into skills. What you demoed for me looks very much like the process I go through to create custom GPT. And like you said, like you, you did a great job like mimicking what the experience is for a creator of a custom GPT. How I where my disconnect is how skills layer in there, like the equivalent of what a skill is in Claude.

Kelly Sinclair [00:38:20]:
It’s just that the way a skill is in Claude is, is it’s about how it’s accessed, right? It’s like its ability to always be part of the conversation. So we’re definitely getting to the next level in the conversation when you ask that. But just even within the window so of like that chat itself so that tool itself self can always access all of the bots that are in there if you like. As we start growing and you start having more bots and more squads for yourself, you can be tapping those in in that it’s just a different way that it shows up because they have like a visual.

Monica Froese [00:38:58]:
So is will it be intuitive like a skill is in Claude to know to tap in or do we have to call it in?

Kelly Sinclair [00:39:07]:
For now we’re working on that level of orchestration. That’s another like programming challenge to get so. Okay, getting nerdy. Get nerdy. If this is over your heads, don’t worry about it for now. You can program certain things to happen in the order like I showed you. And most of the time it usually works where we’re still like finessing it. But the other thing is that there’s an orchestrator, which is like an invisible thing, which is exactly how quad skills work.

Kelly Sinclair [00:39:39]:
That is like watching and going, what do we need here? Where do we need to use this? Who’s best equipped for this? So we have built a different type of orchestration, which is just a thing that you toggle on essentially at your settings level to be like inside of a squad.

Monica Froese [00:39:56]:
Okay, so the, basically the demonstration you just showed us is the equivalent of custom GPT chain linking, but happening without any manual intervention. Essentially like the, the bots know when to pick up where the other bot left off. And that is what custom GPTs can’t do. And if people are wondering, well, why don’t you just build one bot then that does everything, I’m going to answer that because I feel like that is a big disconnect that a lot of people have. When you put too much into one bot, it will go rogue. So like my bots in my programs, for example, are very highly specialized bots that like, like I was saying when I went through the list, what an opt in purpose is, is different than the order bumps purpose, that’s different than the actual digital product, that’s different than building the funnel pages. And so if I put it all in a bot will go, it would be the equivalent of just open a chat and try to figure it out without having the guided instructions, because that’s what’s going to happen. So where as a product creator, as a course creator who was using these bots to increase the value to my students to get to the finish line faster, I was creating like basically these individual bots to help accomplish all these tasks that the program’s helping you do.

Monica Froese [00:41:17]:
And I had, I basically have a whole module that walks you through. Okay, well when you’re done with that bot now, I need you to take the output from that bot and move it over to the to kick off the next. And we just watched Kelly do is the bots just keep rolling. So from a customer perspective, if you were in my program and I was using this, in essence, you should get through the entire system more seamless and faster.

Kelly Sinclair [00:41:46]:
Because like, I know that on our platform roadmap, like I can’t even believe I’m saying this out loud, but Andrew, my business partner, he’s like, I’m jealous of Claude. Like, you know, we should be able to do that. I’m like, dude, we are two people in basement suites in Cochrane, Alberta. You have a day job and they have hundreds of billions of dollars.

Monica Froese [00:42:07]:
Oh, man. Right.

Kelly Sinclair [00:42:10]:
So, like, let’s just cool our jets for a second and be like, how, how do we like connect these things together? Like, anthropic is anthropic. It’s awesome. It does a great job at what it does and what we’re solving. Mostly like we’re solving this multi bot orchestration thing. That is also what Cloud Skills solves. But we also make it sellable. Right. So I think that’s the main thing is that if you’re thinking about it as a product, and this is why we’re here at AI that sells, but thinking about how to use this as a product, how we can integrate, how it can operate like a skill in a way, how it can have a more of a closed container on all the steps, know what all the steps are.

Monica Froese [00:42:50]:
That’s.

Kelly Sinclair [00:42:51]:
That’s what it does mostly differently.

Monica Froese [00:42:53]:
Yeah. So right now it’s solving the specific pain point it would solve for me as the course creator who has the existing nine bots that I want to work together as it solves that chain linking, now the next level. So once skills essentially called in, it really just up levels my bots for sure. But the other issue as a creator that Kelly’s trying to solve is the fact that I could give you a custom GPT and you don’t get my backend instructions or my knowledge files, which essentially keeps my IP protected. Right. When and if you were in my presentation, when I hand it you my skills, you can read the back end of the skill. You can read all my reference documents. And you know, I don’t, I honestly, I don’t have a final formalized opinion of how I feel about that.

Monica Froese [00:43:41]:
There’s definitely some things I would not sell in my skills. Like that is just too, too close to my IP to just give it away or it have to be. I’d have to charge a lot of money for it.

Kelly Sinclair [00:43:52]:
Yeah.

Monica Froese [00:43:53]:
So this keeps all of that protected, essentially. That’s the other thing that you’re solving for a course creator like me.

Kelly Sinclair [00:44:00]:
Yes, absolutely. And one thing I will say too is just a couple of examples of conversations I’ve had with our betas like the AI, just in general, the advancement of AI and how much people are Using AI is going to, it’s changing this course creation landscape. Right. That is not being dramatic, that is just facts. That means that people are literally saying to people like one of my beta clients, oh, I don’t need to be in your membership. I’m just going to take your workbooks and put them in my AI and do them that way. And it’s like, excuse, excuse me, first of all, why are people doing that? But that’s a real thing.

Monica Froese [00:44:39]:
It is.

Kelly Sinclair [00:44:40]:
And so now it’s making her think, she’s like, I’m going to get rid of the workbooks and I’m going to make them interactive bots. So you still get the experience, but you don’t have the.

Monica Froese [00:44:50]:
Well, this is what I’m announcing tomorrow is I’m moving past the self paced course model into, okay, now we’re going to implement it together. And the whole membership is structured on implementation because. Exactly what you just said. This is also why I don’t make my courses downloadable. The workbooks are, the spreadsheets are. But like, I don’t make downloadable transcripts. You, you have closed captioning. You can search and searchable on my platform because of this reason.

Monica Froese [00:45:18]:
Like, it’s definitely an issue. It’s definitely an issue.

Kelly Sinclair [00:45:23]:
Issue and an opportunity. Right. Like you said, you’re trans, trans like translating or whatever, converting your program. And that was within, within 24 hours of me delivering the bot squad for my first White Glove beta client. She said, I think this is going to change the entire pace of my program. The way I normally do it is in like sprints and we meet once a week for six weeks. I think we can do it way faster and we can start doing different things.

Monica Froese [00:45:50]:
I mean, the alone got people over the finish line for me because I’ve been running that program since 2020 and it was all, it was all manual lifting. AI wasn’t a thing. AI wasn’t helping us. And I would have four month sprint containers, four months to get people to get enough data from an idea, an abstract idea to the product’s done, create it launched and now I have enough data that we can read the funnel data in that four month container. Yeah, I mean the custom GPTs cut that down to. I can get people over the finish line. I mean, because we have to give it a little bit of time for sales to come in. In maybe 30 days.

Kelly Sinclair [00:46:25]:
Yeah.

Monica Froese [00:46:26]:
Which is astounding. But like the, all the, the gaps I was having of people getting stuck, the GPT has really helped solve a lot of that.

Kelly Sinclair [00:46:34]:
While the platform is definitely got the structure there to be an entire AI operating system for your business. Like, that’s me saying we’re competing with Anthropic and chatgpt, which is like, right now.

Monica Froese [00:46:49]:
Yeah. Well, hence why I’m moving towards the implementation membership model. Because things are changing so fast that keeping a static course updated for the changes is near impossible. I need a more dynamic format to be like, look what just came out yesterday. Look how what I showed you yesterday. Now, like, Claude just launched Design. So I was about to record my artifact module for the lab and I press. I paused it because I’m like, I think Design’s actually going to take it to the next level and I don’t want to have to record it three more times, you know, so it’s just

Kelly Sinclair [00:47:24]:
like, this is why I just built the software instead of trying to teach people how to use it.

Monica Froese [00:47:30]:
Kelly, thank you so much for this demo. I’m going to drop her link one more time. Of course, the replay and her AI tool launch playbook will be in the classroom.

Kelly Sinclair [00:47:39]:
Absolutely. I want to hear from you all. It’s. It’s a wonderful adventure to build this stuff, but it doesn’t matter if it’s just my ideas. I need other people. This is why we’re, we, we’re letting the people in, we’re getting the feedback and it’s real and in the world. And I’m so excited and thank you for allowing me this space to, like, jump off of.

Monica Froese [00:47:57]:
Yeah, I, you know, I’m, I’m super excited. Since you told me about it, I’ve been super excited. All right, thanks, Kelly.

Kelly Sinclair [00:48:04]:
Thank you. Bye. Bye.