AI engineer and product builder behind Cold Visit, a tool that sends an AI agent to test and audit websites like a real user, exposing friction and positioning issues before they cost conversions

Thanks. So that's why. Yeah, I think you're, I think you're right. Cool. Okay. Yeah. Uh- All right, let's go. [laughs] Yeah. [laughs] So, uh, it's great to meet you today, Raihan. Thanks for taking a minute to hop on the call. Um, yeah. We've just been talk- Absolutely. Thank you. We've just been talking a little bit about what you've been working on with Cold Visit- Yes ... and, uh, kind of what, uh, life and, and tech and AI and everything like that. And you're, you're actually... You've been an AI engineer for quite some time, haven't you? Yeah. About four years now. Yeah. Going to be four years. Yeah. So, um- And what got you into that? Go ahead. Yeah. So I'll start with that. So initially back in twenty-twenty-two or twenty-twenty-one, I don't have the exact timeline. I might be a bit, uh, rusted on the timelines, but when ChatGPT came out for the very first time, everybody was very, you know, new to it, very surprised, and they were very, um... It was very shocking to see the things that it would do, right? Along with ChatGPT, something else came out as well, but that nobody really explored to that extent, and that was the ChatGPT API. OpenAI APIs, right? Today, we know that OpenAI APIs are very public. Everybody's using it, and they're building their own products. But back then, when the models' name were text-davinci zero zero two, zero zero three, uh, Ada, all of these models, nobody used to use it. And I was just playing around, and they used to give this five dollar credit to every new account that used to sign up. So I was just playing around, and I was very fascinated that I can build something of like ChatGPT, and I can, you know, practically you can build an application that can practically do anything that you want. Absolutely anything. It's like an Aladdin's Chirag or something like that, right? So I ended up building an application which went viral online. Uh, it had over-- It still has about four million, uh, active users every month. It's a free nonprofit application I pay from my, my pocket. It's a religious application. Nothing to do with anything, but yeah. And then I later on went on to build a few more applications. Some of them again went viral. I was featured on Abu Dhabi's national newspaper. I was featured on, uh, the Scientific American Journal. It's a, it's a US-based journal, I guess. Mm-hmm. I've been featured in China and Italy's news as well. I mean, I've been, I've been interviewed across multiple newspapers and news channels in the past. And I've been in this space for long enough, I would say, right? And I've been, uh, constantly building, and I have this passion for products where I build solutions to problems that I experience or I see people around me experiencing. And my goal is always very simple. I know how to build. I wanna build something that people can use it without knowing how it functions, without knowing what to do with it, or e- or essentially, I'm an engineer, my job is to build stuff, right? So I've built a lot of projects in the, in the past as well. Right now I'm working at a UK-based fintech. We're building an, uh, AI CFO and planner there. It's for early-stage startups who are going for fundraising and all of that. You can connect your bank accounts there, you can analyze, simulate scenarios, what happens if you hire a new freelancer or an employee, what are the taxes that you'll have to pay, and, and other kind of scenarios, right? So that's it. But, uh, if you ask me, uh, what inspired me to build Cold Visit, right? The product that I'm building right now, is basically, I have built so many products, and I'll be honest with you, I mean, nobody kinda... I don't know why people don't wanna admit the failures. I, I would happily do it because if the failures weren't, then I wouldn't be where I am today. I have built lots and lots of projects, and ne- not every, every single one of them was successful. I think what, ten percent of my projects went-- wa- was successful or went viral at all, right? Ninety percent were failure. Mm-hmm. And not because the idea was bad or maybe the, even the idea was... I, I don't know. I mean... Or maybe something else went wrong or some... I just did not want to pursue or something like that. But, but most... And, and, you know, the place that I work at, I'm not just an engineer there, I'm also a product manager. Mm-hmm. So I am solely responsible for the product that we're building, the, the part of the product that I take care of. So from a product manager's lens, if you see, and from a customer and even a founder's lens, if you see, great ideas and startups and products don't die because of the engineering or the product or even the idea. It dies because of distribution. It dies because it's not clear what they do. Mm-hmm. Right? You go to a website, you see it says, uh, "Your sales pipeline automated in thirty minutes." Okay, fair enough. You're telling me that you're gonna automate my sales pipeline. How? Mm-hmm. There is no how-to link. They want me to sign up. They want me to add my credit card. They want me to invite three people, only then I can see what they're gonna do to, they're gonna do to make my sales pipeline better. There's a wall there. There's a friction. The friction is that they're asking me to put my credit card and trust them before I actually see any value- Mm-hmm ... before I actually even see if it's gonna be useful for me. Mm-hmm. There was this product I was... website, uh, on, on Reddit, I read this, and I, I do... Since I'm building Cold Visit. Cold Visit is essentially, by the way, for everyone e- who does not know, Cold Visit is a platform, is a tool that sends an AI agent to your website, to your product. The AI agent will open it in a real browser. It will go through your website. It will analyze just like a human does. It reads it through. It tries to understand what you do. Then it will try to find the button to do next, whatever it, what you, whatever you want it to do, CTA. And then I'm also rolling out a new feature where this AI agent will sign up to your pr- product using a temporary email address. It will sign up to your product, and it will do one particular action that you define as the first value moment. Mm-hmm. Let's say you are a platform that lets people, uh, plan trips using AI, right? Mm-hmm. So the first value would be you would want a user to come to your platform, create an account, and create, generate their first itinerary using your product, using your tool, right? Mm-hmm. That's your first value. Before they do anything else, you would want them to at least generate an itinerary and see what value you provide as a tool. So- My AI agent will open your product on a real browser, sign up, go to your dashboard, figure out what to do, what buttons do what, and then it will create the first item A. And that has to be defined by you, by the way. So it's a user-defined first mo- value moment, and then it will come back with a very detailed report and tell you where it felt frictions, what could have been done better, and if it reached that first value moment or not, and if it was worth it, and as a user, if it would be willing to pay for it or not. Mm-hmm. So you see, it's like a cold visitor coming to your platform, figuring out everything, and wanting to give you honest feedback of what happened and where it hesitated, and if it is worth it, and what it would want to see improved, right? Mm-hmm. So that is what essentially Cold Visit does, right? Mm-hmm. And the motive behind building Cold Visit was when you yourself, for example, Mark, if you're building a product and you spend one month building that product, every single day you're looking at your product, you're looking at the same website, things will start looking perfect to you. Everything will look- Yeah. Because you're used to it. You will think that you are solving the right problem and you're solving it in the right way. But it might not be true for others, and that is why sometimes you want another pair of eyes to look at the pro- to give an honest feedback, and that's where Cold Visit comes in. It's a one-time report kind of thing. You can run it on your website. You can, you can, uh... This new feature that I'm rolling out probably this week or next week, where it-- you can, uh, on- let it onboard or sign up to your product and try out and tell you what felt right, what did not feel right, and all of those things. So that was the motive behind or inspiration rather behind building Cold Visit. Mm-hmm. That's really interesting. Yeah, that familiarity bias I think really affects just like whole teams. It's not just like one person- Exactly ... it affects everybody. Like, why is the sink there? And it's like, well, we never actually asked that question. Exactly. I've, I've come across a lot of-- I, I, I'll be honest with you, I've come across a lot of products where I've seen that the product are not doing great because not of the idea, the idea is great, but because of the positioning and clarity. The website does not exactly, uh, indicate what the product does or it just does not bring value right away to the user, right? For example, there was this one Reddit comment of a user who had this-- they're building something. And I saw on the website that they are building some sort of sales automation pipeline, uh, or, or some other thing I don't remember, and very good, and they had all the pricing and everything, charging forty-nine dollar or seventy dollars a month and everything. Like, the one and only question that I had was, what integrations do you connect to? Do you connect with my HubSpot? Would you connect with my Salesforce pipeline? I need to know if I'm paying seventy-nine dollars and then after I pay and then you tell me, "Okay, sorry, I cannot connect to this, I cannot connect to that," you know, then it's useless for me. Yeah. So these, this, these are the places where they lack a-and people don't understand. So that's why-- And I, I actually manually, uh, reviewed their website and, uh, reported them back. But gen-- essentially Cold Visit is supposed to do this. So Cold Visit will do that too as well. So if it'll, it'll analyze like the product positioning and compare it to, okay, you're a- Exactly, yeah. Oh, interesting. Not just product positioning. Product positioning, uh, uh, the clarity if, if you're claiming to do something and if it's clear or not, the SEO of your website if, uh, it's, it's probably-- it's properly SEO functional or not. Does it have all the alt images texts? And if it's gonna rank in Google searches. The design review. For example, this is one of the firm things that I've observed in the fintech industry. When you're building something in the finance industry, do not go all black with the theme. Mm-hmm. Right? When you're dealing with money, you wanna be very white and very straightforward with them. You wanna be very, you know, you know... Dark, dark theme websites do not look good when you're building something on the fin- in the finance industry. So these kind of theme reviews. Uh, okay, okay, your website looks a little, uh, sketchy. Your website's design does not

Look good, try to simplify it. ColdVisit audits and reviews back everything and gives a very, very detailed report. Oh, interesting. And what gave you the idea to, to make this, this kind of product that it's kind of like, I don't know what you-- I don't know what the, the one-liner would be for it. It's not like a product manager or product analysis, but, uh, it's, it's very clear what ColdVisit does. But what gave you the idea to create this kind of product? It's because I myself have built a lot of products and have failed, and not because of the ideas. I've been confident. I've, I built, uh, I built a pro- product. Uh, it was an open source thing called AgentNet. It was a really good concept. Back then, there was nothing ca- uh, as MCP or even A2A, and AgentNet was essentially a network of agents openly communicating with each other. Uh, but I did not pursue that product. It was just an open source GitHub com-- uh, pro-- repo. One month later, Google announces A2A, exactly everything as same as AgentNet. Hmm. If I would have just gone to the extent of patenting it, I would have probably got a good amount of royalty fees. But anyways, what I'm trying to say is a lot of products fail because of these distribution funnel pro-- uh, value, not showing enough value at the very beginning or positioning and clarity. And even sometimes I've seen, uh, startups spending millions of dollars in a product, but then when they finally launch it, the numbers are very ugly. It's not because of the product or the functioning or the engineering. It's because users do not really understand what to do. Mm-hmm. Or there wa- there was in-just enough friction for them to convert. Mm-hmm. Yeah, I see. I- So that's how the motivation came. Yeah. Almost like so a lot of marketers use your product then, yeah? Or- Marketers, early stage founders, indie devs people. Right now, I mean, so there's this very c- uh, very interesting trend going on with these AI tools. Indie devs are building a lot of products now. Solo developer going out, building, pushing, launching, and everything. So this is for them- Mm-hmm ... essentially. Yeah. Early stage founders who are launching and, and are not very clear. They are getting good traffic. Every month, hundred k people are visiting their website, but barely anybody's signing up. Yeah. Yeah, it can be very- So that's- ... frustrating trying to figure out what the, what the problem is. [chuckles] Exactly. Yeah. I saw- It lets you download a report in a PDF format. You can share it with your team as well, so you know, and, and you can make it public for other people to look at it. So it's quite a lot of features, I would essentially say. Mm-hmm. A very good detailed report. Yeah. And it's a good website, a good-looking website too. I used it the other day, and I, I, uh, I'm gonna- Yeah ... I fired it at the HiRey website. I'm gonna use it some more just to give it a, give it a good run. But yeah, very high quality. I saw one case study where it was like a, a travel company similar to your airline, uh, airline metaphor earlier. Um, maybe it was TripAdvisor, maybe it was Expedia, but it was one field on their form, uh, that was causing it, it caused them to lose like millions and millions of dollars of conversions because one field- It happens, happens so much ... on their form was too much. Yeah. Happens so much. Some companies just think that they should take, uh, almost every information or during sign up, and that causes a lot of users to drop off. 'Cause once you ask the users to just give twenty field informations, that's where they're gonna drop off. So there was this one audit that I was doing. Actually, the agent, uh, on ColdVisit did. Uh, ColdVisit tried to onboard itself to a platform, sign up to a product, and what happened was that there was no, uh, red asterisk mark on the agency name. That means the agency name was not a required field. But when it tried to sign up, it threw an error that you need to put an agency name. So it reported back that as well. Mm-hmm. Mm-hmm. Yeah, that can be very frustrating too, not knowing whether something's required or is this required or not required? I had a lot of conversations, uh, with different stakeholders about that. Yeah. Yeah. Oh, cool. So where do you-- So where-- So you're adding these, these next few features to ColdVisit. Do you think you're gonna keep on working on ColdVisit for a while, or do you think you're gonna pivot it, uh, make a n- a new, uh, another type of product? Or where do you see, uh, your, your personal and, and, uh, product pipeline over the next, uh, few months or a year? I'm gonna go-- I'm definitely gonna work on ColdVisit more. I'm gonna give it at least, at least three to four months before I actually turn on my eyes on it, if at all things do not work out, but I'm still gonna work on it. Uh, these few features are gonna come out this week or at max probably next week, I'm gonna roll it out. It's almost ready. Mm-hmm. I'm just wanting to do some early testing with some users before I finally roll it out. And I am entering into the direction of testing, right? So I could probably enter into the QA testing phase as well. Mm-hmm. I'm thinking of having some sort of recurring testing thing on the platform where you can configure with your website or code base, and then it regularly tests, uh, sorry, uh, it, it, you know, on a recurring basis, it tests your website and reports back with insights and all of those things. So yeah, interesting direction. Let's see. Mm-hmm. Still exploring. Yeah. Yeah. I'd love, I'd love to test out, test out, uh, new features on it and, uh, kind of see how you develop it. And what, um- Okay ... where did your first few users come from for ColdVisit? Where did your first, your seed users? So if I really be honest, uh, so first few users came from Reddit, to be really honest. Yeah. Uh, yeah. Reddit was the thing. But the first two paid users came from, I would say, personal connections only. Yeah. And by personal connections, I don't mean I'm forcing people to pay. I just, I know-- So I have in my connections, I know a lot of people are building stuff, right? So I just share with them casually, tell them what it does and how it does it. And by default, they were interested to try it out. And they tried it out, and they bo-bought the, uh, they unlocked the report, they paid for it, and they really liked it, an- as an honest feedback. And I've even asked them to give a testimonial so that I can put it up on the website as well. So first few users came from Reddit, but first few paid customers came from my personal, uh, network, I would say. Ah, yeah. Yeah, that, that makes sense. Yeah. Personal networks, I feel like a lot of people, uh, tap into those when they're starting a new product. Yeah. Exactly. Yeah. Reddit is kind of filled with people trying to get each other to pay for stuff, but for, for just, you know, uh, getting people to try it out, it's fine. But I've hardly seen anybody convert from there, including yourself, Mark. Yeah. [chuckles] Yeah. Yeah. I just, uh, I was-- I've been on Reddit, uh, doing, getting a lot of, uh, early stage founders and this type of thing on, uh, you know, te-testing it, interviewing them, getting them on our platform and everything like that. But I will be using your product. Yeah. Yes. Cool, cool. Yeah. All right. Well, cool, Raihan. Well, I gotta get to another interview, but, uh, thank you so much for taking the time to, uh, to talk with me today about ColdVisit and about your background and about, uh, what you're gonna be doing, uh- Definitely ... in the future. Yeah. It was really gr- great talking to you, Mark. Yeah. And I think that's it probably. You have a- another call, I think. Yeah. Yeah. Yeah. And I hope you unlock your full report so you can give me honest feedback because I'm still waiting for it- I will ... from your side. I will. You got it. And I hope you, uh, you're able to create an account on, uh, the HiRey, the hirey.ai network t- as well. Yeah, yeah. I checked it out. Yep. And I'll- I checked it out ... we'll, I'll take a clip or two from this interview, and we'll upload that to the website. That'll be associated with your profile. And we'll-- We're gonna do different feature pages with different, uh, founders and products like yours. So as traffic comes to our website, it'll, uh, we'll, we'll hopefully get you some more users for ColdVisit, or if there's any connections in HiRey that fit you well- Exactly ... we'll, we'll recommend those. Okay? That would be awesome. Thank you so much. You got it. Thank you so much. All right. Take care, Raihan. Bye. We'll talk to you soon. Yeah. Bye. Bye.

Captions (.vtt) · esCaptions (.vtt) · jaCaptions (.vtt) · koCaptions (.vtt)Captions (.vtt) · zh