Build an App & Get Clients in 45 Days (No-Code Claude Tutorial) I Ft. Igor Krysnakh

Speaker A: And within one week. So like we're talking five business days. I've created the entire front end for that application for the pivot. But you just need Claude at this point. They told me it will take roughly around nine months to build it out. Nine months. I was like, that's bullshit.

Speaker B: Not nine months. Five days.

Speaker A: Yes.

Speaker B: That's pretty damn cool.

Speaker A: You know, this thing saved me a lot of time, like months worth of development.

Speaker B: That almost feels like we're getting to artificial general intelligence. Privacy is potentially an issue if you're giving these agents they accidentally leaked 500,000 lines of code. I had tried a few times, but I just wasn't getting the output that I want. But I realized that's probably my fault.

Speaker A: Uh, they never anticipated such growth. And when they opened that product to the market, they were assuming that only a few people will use it. And then they started to have server problems, servers crashing. So they had to scale faster than any other startup. If you do nothing right now, then your business will lose this much money by doing nothing.

Speaker B: Okay, I want to create my own company. I want my own empire. In 45 days. What are the different components that they need to stitch together?

Speaker A: When you're building a page, you can write a humongous prompt explaining in very explicit detail what you want and where and how you want things to operate. We take that file from chat, load it into cloud code, and say, here's the instructions.

Speaker B: And he said that recently he's not writing code at all.

Speaker A: Historically, you had to have a whole engineering team, some certified people who can come in and help you figure out how to do it.

Speaker B: Today we're going to be discussing the 45 Day Empire with my friend Igor Krasny. Really excited to have you here, Igor. So tell us a little bit about why the future of business is not working for somebody else. It's about creating your own thing.

Speaker A: So if you guys see what's happening with AI these days, AI is revolutionizing how tech industry is evolving. And it is becoming much faster for people to create new businesses and, uh, make their ideas come true using AI than it used to take like 10 years ago. And my projection, or what I think will happen soon, and what you will see happening soon is people will be coming up with ideas like they used to and they will be implementing, uh, those ideas at a much faster pace. Therefore, we will see more entrepreneurship happening these days and you will see more smaller businesses coming or surfacing. With AI being where it's at and how easy it is to create new software, I Believe that more and more people will take advantage of AI uh and will be starting new businesses. They will be doing at a much faster pace and they will be creating seven figure, eight figure companies by themselves or with a small team of three to five people. And if I were to give advice to my son right now would be to invest into that creativity outside of the box, thinking and making sure that he uh, is able to identify opportunities in the world and apply, whether it's AI and his knowledge to build something out of it.

Speaker B: So you have a background as a technical founder, you were in the enterprise space and then in 2019 you became a full time entrepreneur. Building multiple businesses, having failures with a couple, having great success with their few. But now it seems like the moat or the time to an MVP is collapsing.

Speaker A: Yes.

Speaker B: And you're using Claude code, correct? Define this for us.

Speaker A: So, well, let's take a step back. Um, I'm sure many of you guys watching are familiar with what cloud code is, but if you're not, uh, there's a company, Anthropic, they're a competitor to OpenAI or ChatGPT product. Uh, anthropic has a product called Claude and within Claude you have different features like Claude Chat, Claude Cowork and Claude Claude code. And they're built for different purposes.

Speaker B: Not to be confused with Open Claw just to further complicate things.

Speaker A: Right, yes. So, um, let's come back to the mode. Uh, yes, I do see things have accelerated quite a bit. And let's maybe back up a little bit. As I was growing up, I think uh, I first developed a passion towards front end technologies, HTML, JavaScript, CSS when I was in high school and I used uh, text editors to build websites. Back then, uh, fast forward a little bit, you had other tools like some people use Dreamweaver, some people used, I don't know, other tools to build websites. And some of the elite front end programmers still used web editors or kind of like IDEs, right. To code HTML, CSS, et cetera.

Speaker B: Integrated developer environments.

Speaker A: Yes. Okay, uh, where we are right now and I feel like we're kind of jumping over a huge portion of huge time frame here. But uh, here's. This is the story of where things are right now. Um, if we slightly rewind back maybe a year ago, um, that was the first time where I've heard people are using ChatGPT to create basic landing pages. I've even seen some M. AI startups with agents where you can go and say, hey agent, so, so and so build me, uh, I have a photo, uh, business build me a website and the agent goes, cranks out HTML for you. Very basic. Uh, that was about a year ago now where we are today. And this is kind of where my experience comes in. Uh, at about uh, when was it? December last year. I made a very big pivot with one of my businesses and that pivot required rebuilding an entire front end for the application, changing the website, et cetera. Um, at the time I had a small team that was working with me on that product and uh, those engineers were backend engineers. They didn't have a lot of front end expertise. Whenever we uh, had conversations about adding features, they always kind of uh, shrugged their shoulders in terms of how long front end features would take to implement and uh, it was always a bottleneck. So uh, after we made that pivot in December, I ended up downloading cloud code and was just trying to kind of figure out can I translate or transform the uh, wireframes that I created for my dev team previously into fully funct prototype that they can wire the backend into. Um, that's where it all started. Um, it took me about a week to get familiar with cloud code, but within a week time frame I was able to create an entire end to end front end prototype with all the features that I wanted. Not only for MVP phase that we were discussing with my team but literally like a year ahead. Because their original estimate, when I showed them the wireframes they told me it will take roughly around nine months to build it out. Nine months. Given that I had a uh, or have front end expertise, I was like that's bullshit, you know, that should be faster. But I didn't like, I was thinking like maybe a month or something like that, maybe a couple of months to build the front end and wire the backend will take longer. But I took cloud code, started using prompt engineering and just start feeding it wireframes and optimizing those. The result that cloud code was giving me and within one week, so like we're talking five business days, I've created the ENT front end for that application for the pivot.

Speaker B: Wow.

Speaker A: So we're talking like again in my previous estimation was maybe a couple of months. I've already shrunk it down to five

Speaker B: days including the design process.

Speaker A: It's design ui. Like just think of it as a face of the website. HTML, CSS, JavaScript, like the basic stuff, multi page search connected using mock data. Right? Because like we don't have a database at that point. So just an entire experience using Mock data was done within five days.

Speaker B: So. And now you're innovating the E commerce space, which is a, uh, 1.5 roughly trillion dollar market in the US every year and accounts for 20% of transactions for retailers. But we'll get into that in a little bit. But if somebody is watching this right now and says, okay, I want to create my own company, I want my own Empire in 45 days, what are the different components that they need to stitch together? Do they need a Mac Mini? Do they need to download Claude code? How does, how does it work?

Speaker A: Here's where I would start. You can get a basic or. I don't even know if, uh, CLAUDE has a free version right now because I'm on a paid subscription. I haven't even looked at the free version at this point, but I think initially I got a free version. Then very quickly I upgraded to a $20 or $25 per month plan because I was running out of tokens. Um, just to see, will this give me what I need? Can I, um, unblock my backend team to be able to continue building this product onward? So first step in terms of your hardware, Laptop, you don't need anything else. Everyone has a laptop these days. Uh, you just need laptop and access, uh, to Claude. And within Claude itself, you do have cowork and code and chat. We can talk about those a little bit, uh, separately, but you just need Claude at this point. So Claude and then when you use Claude code. So since we're jumping into Claude code, um, I connected it to, uh, a version control system, which is your GitHub or GitLab. We're using, uh, GitHub, uh, within my company. So it just, I created a repo, connected my report, told cloud code, here's where stuff is. This is where you can put the code and then it's from there, it's all prompting.

Speaker B: Are there any specific concerns? Because obviously privacy is potentially an issue if you're giving these agents, because cloud code is essentially an agentic ecosystem, a framework. And if you're giving access to your private data and your applications in your emails, like, this could be concerning. So did you use Docker to create like a safe space within your computer or how did you approach the security concerns?

Speaker A: I didn't, because when, you know, I had ideas, I was just trying to say, will this product help me build this faster or help me get to my finish line faster? Um, but at the point after I did proof of concept and said, all right, you know, this thing saved me a lot of time, like months worth of development. And I knew that my team will have to adopt this product even for the backend purposes when I knew my entire team was going to be using it. We've looked at the enterprise or higher paying plan within claude and what I've, uh, the small research that I've done is when you're on a free or $25 per month version, then I think Claude gives you a uh, message there that whatever you put in there, they will train their LLMs based on that data. So security is not there. Right. But when you are on that higher tier plan, enterprise plan, they say they don't train you uh, on your data, which kind of gave me the confidence, so to speak, to go down that path.

Speaker B: It is a little ironic, speaking of security, uh, and anthropic, given just, I think it was two days ago they accidentally leaked 500,000 lines of code with a configuration file that was leaked. A map, I think it was.

Speaker A: Okay, I didn't hear about that. So.

Speaker B: Yeah, yeah, yeah, that's.

Speaker A: Well, I mean, look, they're a fast growing company. I think when you look back at uh, the OpenAI or ChatGPT story, uh, they never anticipated such growth. And when they opened that product to the market, they were assuming that only a few people will use it. And then they started to have server problems, servers crashing, so they had to scale faster than any other startup prior to them. So again with Anthropic, will they have issues? Of course. Um, and, but uh, you know, at the end of the day, I believe that given where CLAUDE is and the tools that it provides or gives startups like mine, for example, to use and leverage is worth the risk at this point. And you know, another thing that I will add is Claude, uh, code is used not by only small businesses. So I know Facebook is using cloud code. There's a bunch of other companies that do have enterprise access. So I'm pretty sure CLAUDE is on top of their security game from that perspective, as much as they can be.

Speaker B: When you decided to build a company, what was the first place you started? Did you go to claudcowork.com and download an installer? Or how does somebody begin if they want to create a company for themselves?

Speaker A: Sure. Well, it all starts with an idea, right? Like you think that you have something of value that other people in the world will benefit from. Right. And to me, I've mentioned the pivot that I've done so before. I'm going to take a couple of steps back before, uh, the company that Kind of evolved into this new product was a data synchronization platform. So taking E commerce data, taking CRM data and moving the data or synchronizing the data between two systems so that no matter which system you look at, the data is accurate. Right. The pivot that I've done is I see that with AI people or companies creating data lakes, slapping AI on top of it, the need for the integration is kind of dissolving if that's the right word. It's not as high of a need these days and things are shifting more into business analytics or system analytics, uh, leveraging the data that you have to be able to do more with it. Right. So anyways, I'm going back to your question. How do you start? You start with an idea. And um, that's probably a whole story on its own how I came up with this idea. But I knew that I'm going to build a platform for businesses, for merchants to highlight what's m important to them. Like think within business. So much stuff is happening day to day, every minute and you don't know what's important, what's not important, how to prioritize. There's a lot of noise. And the platform that I'm building, it's a business signals platform. Those events that occur, I call them signals and I surface the critical signals, important signals at the right time to those merchants so they know exactly what to act on to maximize their revenue growth and minimizing their potential loss within the business. So from that idea I went to cloud code and I've done like brainstorming sessions, research, I don't know what do you want to call it, but it was probably many hours of back and forth conversations with uh, cloud chat on how do you take an idea and what type of like what needs to be implemented. And uh, that kind of back and forth conversation was more about based on my personal expertise and business within the E commerce space and the questions or challenges that uh, my customers, merchants were bringing to me over the past two decades. All that experience was just with cloud code brainstorming, figuring out what's critical, why Claude was sometimes hallucinating and I had to put it in the right direction. But uh, to keep this story kind of short and on point is it's back and forth conversation until you have an idea or a plan that you can go start to implement. Then from that idea the next step for me was to take a look at the wireframes or create some sort of a UI that people could use to access this idea. And from that ui, the next step was to actually getting into cloud code to create the interface for the ui. Like and ah, those are kind of the initial steps.

Speaker B: So would you say claude code is like a terminal based powerhouse or are you actually in the command line interface? Are you talking to a, how do you interact with this platform?

Speaker A: It's a bot basically, just like ChatGPT or cloud code. You have a prompt where you say, hey, I want to build this page. Here's what the header looks like, here's what I want the menu to look like, here's what I want on this page. And uh, kind of one, um, not caveat, but one piece of helpful uh, information that I can offer is when you're building a page, you can write a humongous prompt explaining in very explicit detail what you want and where and how you want things to operate. Um, but I took kind of uh, additive approach with uh, the cloud code and I said, look, build me a template based on this wireframe, black and white wireframe, and instead of doing it in black and white, here's my color palette and brand kind of styles, so to speak, that apply to this template and build it for me and make it look super nice, make it responsive, make sure that it's mobile. First type of a design, go, then it goes, does that thing, gives me back the response and then from there I go, actually don't like how the menu looks here. And you adjust it until you get to the point where the template is done.

Speaker B: This is super fascinating because, you know, you have all these agents, but they have kind of general understandings. But if you can give them skills or specific files that they can tap into, they can be experts in a specific space and teach each other. So skills Designer Sam, and correct me if I'm wrong, are markdown files where you can run a script or a set of instructions. And if you can have these agents learn how to tap into specific information and coordinate with each other and continuously learn or recursively learn. That almost feels like we're getting to artificial general intelligence.

Speaker A: Um, kind of. I mean I wouldn't go probably as far as call it AGI yet, but we're getting close to the point where you don't really need to give very explicit and detailed instruction and worry about hallucinations as much. Uh, so talk about the MD M file. So I use that type of a process every day with uh, my team, right? So like, let's say we need to build a feature or we need to build a new signal that I'm just going to give you an example. Like let's say there is a signal that uh, detects when your uh, VIP customer shops on your website and abandons the cart, uh, what happens? And we basically had to articulate like, how do you identify VIP customer, et cetera, et cetera. So that conversation is done through Claude chat to then build a spec. And the spec could be then translated at the end into that MD file, the list of instructions. And depending on the complexity of the feature, that MD file that we end up creating at the end of that conversation could be like, sometimes it's 10 pages, sometimes it's 30 pages with examples, database queries and so on and so forth. And when we feel like that document is, I want to say, like maybe 70% good enough stuff where we can't pick, uh, or identify any like burning issues and gaps, we take that file from chat, load it into cloud code and say, here's the instructions, help me build this out, give me a plan. Right? And the cloud code goes. Studies your repos. It could be one or multiple repos, depending on your infrastructure. And then it says, all right, I need to modify these files here, I'm going to create unit tests for you and et cetera, et cetera.

Speaker B: When's the last time you actually replace, wrote a line of code? I know Andre Karpathy who is just in a recent video, and if you want to check that video, click this link here. And he had said that recently he's not writing code at all.

Speaker A: Same thing for me. I don't want to kind of put myself at that level. But what I can tell you is, um, probably around 2019 is where I had to make a pivot away from being hands on co code type of uh, individual or developer, engineer, whatever you want to call it. And I went into more like marketing, sales, learning those aspects of the business with my customers. And I think around that time was the last time I wrote a single line of code. Right now it's all like, I don't know what you call it, vibe coding a little bit. But with cloud code it's a little bit more than vibe coding.

Speaker B: More sophisticated.

Speaker A: It's more sophisticated vibe coding right this point. It's been many years since I've actually opened an IDE and like typed a line of code or something like that, or made a fix to an issue.

Speaker B: But you're still creating a product. And not just a product, but a product with paying customers. And you got there from within 45 days?

Speaker A: Yes, I front end, um, uh, aspects of work we're working here. It took me five days roughly to create that entire application with all the features that I wanted that uh, my engineering team previously pushed back on and said like, oh, this is a big effort. We don't want to do it right now.

Speaker B: Nine months, five days.

Speaker A: Yes.

Speaker B: That's pretty damn cool.

Speaker A: Yes.

Speaker B: What is the future of dev teams look like?

Speaker A: I believe that be a new layer or level of engineering that will continue and will evolve. But as an engineer, you still have to think creatively around like, and I'll give you a specific example, we are onboarding our second customer who has uh, about 130,000 orders in there, uh, e commerce platform and they have orders with 400, 600 line items in a single order.

Speaker B: It's a lot of SKUs.

Speaker A: It's a lot of SKUs. Well, and yes, it's a lot of SKUs on the order itself. What we've learned is when we try to onboard that customer and say, all right, let's, let's pull all of their orders into our system. Our system started to make weird noises and slow down and sometimes crash or freeze and we're like holding. Holy crap.

Speaker B: System failure.

Speaker A: Yeah, system failure. So what do we do? All right, so we knew that we needed to do some optimization and the prompts that we used to create the V1 of the product, it built us a beautiful product. It worked in small volumes. But then the problem of scalability and uh, well, yes, horizontal scalability came into play. Right. So at that point we basically had to uh, re engineer our prompts and basically ask Claude to say, all right, well imagine our system, you have thousands of customers with 16 million orders per customer and everyone has 600 line items on the order. And you have to do all these calculations and batching, uh, how do you make sure that the database doesn't fall over, how do you make sure there's horizontal scaling enabled in your infrastructure and all of that stuff. So I want to kind of focus on that for a little bit if you don't mind. Because if we rewind again a couple of years ago, couple, five, 10 years ago for a startup, when you, when you're starting a business, these types of decisions, whether to implement scalability on day one or not is always like the default answer to those decisions or questions was not right now. Because first we need to get some customers on the platform and then get some revenue. And then from there you look to figure out like, well, how do we scale how the AI and Claude specifically the tools that we're using allows us to solve for those problems again within days instead of months without having to know intricacies of infrastructure. Like on my team, I don't have anybody who knows aws, like who has AWS certification. We are hosted on AWS and everything that we've built on aws, we have blue, green deployments, automatic deployments, horizontal scalability based on usage and all of that stuff. And we built it by leveraging AI, expression, expertise, knowledge, and we're doing automatic deployments that way and I can solve for it again within a couple of days versus historically you had to have a whole engineering team, some certified people who can come in and help you figure out how to do it.

Speaker B: You know, what I've been doing recently is taking screenshots because it's, first of all, it's more efficient. Second of all, you know, a picture speaks a thousand words. And because these models can ingest images and video and understand contextually, it's just been kind of a huge unlock. So rather than me having to type everything out into Gemini or whatever, I'll take a screenshot, paste it in there, add a little bit of flavor and it just understands.

Speaker A: Give me an example of some of that. Because I use it also on daily basis. I, um, just wanted to compare a little bit.

Speaker B: So I built a, I built a website probably like three weeks ago, three or four weeks ago. And uh, it was my first time building a website. It was kind of cool to see it, you know, to see it compile and be live, uh, in Anti Gravity, which is Google's integrated developer environment. And there were certain questions, um, that I had in the terminal. So I would take a picture of that, which included its own instructions because Anti Gravity has its own chat as well. Uh, but I would paste it into Gemini just because I'm more comfortable with Gemini. I'm used to it. And in Gemini I'd say, oh, you're almost there, but you need to type this into the command line or you need to update this. You uh, need to access this library. So I was just really just trying to build something in, in this window. And when I didn't understand taking a screenshot, posting it here, getting feedback and taking that feedback and implementing it into, uh, into the ide. And so before you knew it, I mean, it took me probably like a, a couple hours. But I had built, I had built um, a new website which was like really awesome. You know, for me that was like a unique experience.

Speaker A: Where I use those types of screenshots is for example in some cases, our database queries were slow. When I look at my interface for my collection, I don't know how do I type in all the indexes? Am I going to explain it incorrectly? I just grabbed a screenshot of all the indexes that I see on the page, dumped it into cloud code and say, here's all my indexes are indexes

Speaker B: libraries that need to be accessed in order to, for a tool to function correctly. What is it exactly?

Speaker A: Indexes are used for you to be able to do it much faster, like cached structure of data, which allows you to find the records that you need very quickly. The other example is mostly with front end like let's say I don't like how the button looks or the positioning of things or if something is off, I just grab a screenshot and say, this is what my page currently looks like right now. Here's the changes that I want.

Speaker B: Bullet A, bullet B, bullet C. Sometimes it's still struggle. Like I had a uh, graphic that was cut off on the right hand side and I wanted it to extend another probably 15% but I wasn't. I had tried a few times but I just wasn't getting the, the output that I, that I want. But I realized that that's probably my fault.

Speaker A: Another aspect which I really, really like about AI is your application does something but you don't know something is wrong, but you don't know what it is is. And what I often do is grab a file the logs that were output by my application and it could be a large set of data that by hand reviewing it is difficult. Right? But I can just throw that file into Claude and say here's what's in the logs. And in many situations Claude was able to say, oh, this piece, and this piece is working fine. The issue is somewhere here. And by the way, I also noticed this thing in the log analog which doesn't look right. Do you want to also tackle that? And it just uh, highlights in a large blob of text, it highlights things that, you know, surfaces those things to you and then you decide do you act on them right now or later?

Speaker B: Tech arcs yet we mentioned Enterprise access context architecture and MCP connectivity. What are some other elements? From a technical perspective, somebody needs to get going and start building a business using cloud code.

Speaker A: So in a huge kind of couple of things for you to be aware of. So. So Claude has different models like uh, Sonet Optus. Right. Hopefully I'm pronouncing it correctly. Opus 4.5. I think it is Opus and Haiku, I think they have those three, uh, those three different models, they use tokens differently, like deep reasoning versus not right. So haiku is their low level reasoning type of uh, a tool. So depending on your task, you might use haiku for just kind of text brainstorming, et cetera. But if you need deep reasoning, you would go for opus, right? Me is those hallucinations, at least with code, um, come from either poor prompt or maybe lack of context, which would change the direction. So depending on your task, if you can, you know, let's say I want to build this chair. Maybe your prompt should be let's build a leg first, right? And give it to AI. AI will do a pretty good job. And then you kind of, of piece it together with AI and saying like, okay, now that we build a leg, let's build a seat, and so on and so forth. And that is, I think it's a good analogy to use in your kind of day to day interaction with AI because the simpler the tasks are, with a clear output and clear requirements, AI will do a pretty damn good job in my opinion. In the digital space, there was a pretty big moat before, but now it's gone or shrunk quite a bit. Rewind. Ten years ago, when you start a business or launch a company, you would go seek investment, hire, you know, large team, 20, 50, 70 people for you to start building that product. And the more people you have, the faster you could build that product. And then you had a portion of time where you are, quote, unquote, more or less safe until somebody starts crossing that moat, right? And uh, trying to recreate your product with AI. That mode used to be what, 10, seven years? I'm just guesstimating at this point, but now I would say it's like 30, 45 days, probably maybe 60 days in some situations. And that mode to me is, I don't think that mode is static, right? It's never been static. Because when you build that mode, you continue to evolve your product. That mode continues to get larger and larger and how fast you are moving and how fast they're moving. And is that mode decreasing, right, or increasing? Increasing. That's the safety factor. And before that moat was so large, it took years for somebody to, let's say, build another Salesforce or build another HubSpot. But now with AI, with digital tools, this moat is much shorter. Yes. And it's been kind of fascinating when

Speaker B: you look at the SaaS market in North America. It's a couple hundred billion a year and their Tools to make humans more efficient. Efficient. But when you consider that the labor market starts to get eaten away at these tools because the tools are being done by themselves, that starts to change the game in a meaningful way. Right?

Speaker A: Yeah.

Speaker B: Speaking of customers and how you're able to drive a lot of value, help us understand what noise and signal you're able to separate and drive a lot of value for the e commerce space. You worked.

Speaker A: The most common kind of big, uh, issues that occur, they all tie to revenue. Revenue and the platform that we've built and just launched. I only watch two metrics within the platform and surface those metrics to the merchant. One is what is your actual revenue gain from this platform and what is the actual revenue loss? And those two are very important because our goal is to minimize the revenue loss for the merchant and maximize the revenue gain. Gain and revenue loss. And revenue gain are supplemented by two additional kind of, I call them submetrics. It's potential gain and potential loss. So our platform is basically detecting events that can cause you a loss, actual loss. And we then capture that event and say if you do nothing right now, then your business will lose this much money, money by doing nothing. Or here's an opportunity that's sitting here. If you do nothing, then you don't win this deal or this opportunity. And by surfacing those signals, we tell basically or highlight to merchants when to act and what to act on right now, when it's important to them. Instead of waiting like a week or a month later and be able to identify a potential missed opportunity from a dashboard board.

Speaker B: It's fascinating to me about when you mention an event because in reality an event could mean a specific timestamp or a state change of something. An agent could, could uh, impact an event or could trigger an event. Quad code has this, uh, this piece called Heartbeat, which on a consistent recurring basis is able to re, engage or check the state of something and if that has changed, change, take action.

Speaker A: Correct.

Speaker B: So it's truly becoming more autonomous.

Speaker A: This is the next step within automation. Right. It's knowing what are the signals that are occurring within your business, which ones are important to act on. Right. And even with the, uh, customers that we have on the platform right now, we're not just stopping, saying, here's the signal, we're waiting for the next one. Right. Like I, I am on calls with my customers, uh, that are using the platform today. Just saying. All right, so this signal, trust triggered. What do you guys do from here? Is this do? Does your sales team go, takes this event and calls the customer, do they create a ticket? Do they talk to somebody else internally? Like what is that process? Because at the end of the day signal detection is the first phase, I guess in this whole thing then what do you do with that signal is the phase number two. Because in some situations the activity that follows that detection can either be manual or automatic depending on whether or not AI or your automations that you have within your business can handle that signal properly and in the best possible manner.

Speaker B: Is that trigger a behavior? Is it a keyword match?

Speaker A: How uh, does that really depends on the signal. So I'm going to maybe mention a couple of signals. Abandoned cards, right. So usually within Shopify or Magento Adobe E commerce platform you can have uh, extensions or apps that handle your cart abandonment. Meaning when somebody leaves the cart, doesn't place an order, that extension picks up that cart and sends it to Klaviy or HubSpot for email marketing, uh, nurture sequence, whatever. Or you may uh, you know, usually it's email but if it's manual, old and give somebody a call. Right. To that customer depending on the business. But in the most cases I feel like it's either a text sequence, uh, or email sequence. Right. So with the abandoned cart, the big issue is detecting the cart. Uh, one of the fascinating discoveries that I've kind of came across was first of all, there's no abandoned cart view. So merchants are flying blind unless they install a plugin to handle abandoned cart carts. And usually within abandoned cart functionality there's gaps where you could create different flows based on your customer segments and nurture them differently. But that requires you to kind of figure out like, well, which segment needs which nurturing sequence, et cetera. Or many merchants don't even go to that level of sophistication. They just say if it's an abandoned cart, send them a three part email, you forgot something, something and if they don't act, send them a coupon code. Right. But a lot of those abandoned cart sequences don't look at other data which can help you figure out what is the right thing to do. Example, let's say somebody like John Smith goes to your website shops, they wanted to order something for their office for like, I don't know, like a birthday party celebration for their team and they added something to the cart and uh, they get a error message on the checkout saying your card has been declined. Right, right. And they're like oh well let me try a different card. Doesn't work off, they go off of their website. Right. So in that generic sequence you'll say John Smith just tried to buy this product abandoned card. They never placed an order. You forgot you, you get your first John gets first email saying look, you forgot something in the card. John already knows that they forgot something in the card because the card got declined. Whatever, they couldn't buy it. And then eventually they get coupon which again to John would be useless because his card got declined Client. Right. And the missing context around this is what if John is traveling and it was his bank who flagged the card as potential fraud and blocked that transaction. Right. But you don't know and those uh, abandoned card extensions don't know this data.

Speaker B: And uh, you could I guess query the IP address in that, in that

Speaker A: scenario or something like that. Right. So there's different data points, points that we can see that John was traveling in Italy because IP addresses in Italy card declined you from your payment gateway you may get even the specific technical error message saying that suspected fraud. Right. But that information is not easily accessible or serviced to the merchant or to the sales rep in any meaningful way. Regardless whether you're on Shopify or Magento. I haven't seen anything like that. And being able to, to capture that data and say look, uh, Nick, John just tried to buy something from your store. He tried to buy it from Italy. Card got declined because suspected fraud. But you know John because he's your high valued customer. Right. At that point you can send John a text message or an email and say look John, I saw this happened. Uh, the bank declined your card because it was a suspected fraud. But I saved your card. Let me know when you're back or let me know when you lift that uh, blocked from the card. I'll process the transaction on your side. It's a lot faster.

Speaker B: These things add up quite a bit.

Speaker A: Exactly. So that's just one example of the signal. The other one that was just blew my mind because I, I, I didn't think that was going to be at the top of the list. But it came through as the most requested signal from our existing customers is unexpected um, price change. And what I mean by that is if you have a store, most likely you're not manually managing the pricing within catalog. Right. You either have somebody who either does it manually for you or you have another system that where you stores all your pricing and you import it More established businesses price changes may occur on the website, but what if those price, what if the price of the product drops by more than the threshold? So we have one customer who says my threshold for price drops is going to be 10% so any price drops of the product more than 10% I want to be alerted to about or maybe your integration breaks and wipes the product price. So your product is zero all of a sudden and you only find find it or find that problem when orders are starting to come in at zero price. Right?

Speaker B: Sure.

Speaker A: So it's detecting those types of things again. I never thought it would be such a big problem. But for more established businesses it's just that extra layer of protection to make sure that you don't lose revenue you

Speaker B: so somebody has an e commerce site and they want to maximize the revenue. They want to mitigate losses. What's a good way for them to get in touch with you and start to leverage this technology that you've built?

Speaker A: Go to the website idealdata IO do offer 15 day free trial right now uh we'll onboard you requires zero development effort on the merchant side and we'll see set it up for them and they will start receiving those signals pretty much immediately. We're like talking less than 24 hours on boarding.

Speaker B: You're the man. Really appreciate it so much and congratulations again on the success so far.

Speaker A: Appreciate it. Thank you. Thank you.