I Gave AI 7 Days to Get App Users

Is it possible to go from an app idea to your first customer in one week? Well, I'm challenging myself to find out. Last year, I built a software product that went from zero to a million users in 60 days, but that required a full team, months of work, and a six-figure budget. So, for this challenge, [music] I wanted to see what would happen if I stripped it all away. No developers, no marketers, and no personal [music] brand. Just me, AI, and one goal. Take a random app idea and turn it into a real software business by getting my first customer as fast as possible. To start, I needed to find a gap in the market to decide what I should build. And so, I started where anyone in my position would start, and I started looking at what people are currently buying. I went to the App Store and saw that this astrology app had over 200,000 ratings. And after looking into the astrology market, I realized that it's a $12.8 billion industry that's growing 20% year over year. As I scrolled the category, I noticed almost all of them target women, and they primarily focus on relationships. Literally none target men. So, there's only two explanations for this. Either A, men don't pay for astrology and it's a dead market, or B, this gap in the market is my opportunity. And after doing a little bit more digging, it turns out that 18% of men actually believe in astrology, which is about a billion people, including President Ronald Reagan and billionaire banker J.P. Morgan. This gave me the data I needed to feel confident in building the MVP as a web app. When I started building the app, I used Claude Co-work [music] plus Wasp CLI and started describing all of the features that I wanted in the app. Then, I asked Co-work to turn this into a product requirement doc so I could take all of that information and turn it into something that lists everything I want built. I then took that product requirement doc and pasted it into Claude Code, and within just a few minutes, I had a functioning V1. But, it looked mega vibe-coded. So, I spent the next few hours designing the app to make it feel [music] more premium. To do this, I first needed a brand guideline that would make the app feel more masculine. So, I went back to Claude Co-work and asked it to build me a brand kit. I wanted a blend of astrology imagery with colors that implied wealth and in a few minutes, [music] I got custom brand guidelines that defined brand font, colors, and more. Then, I told Claude Code to redesign the V1 using the new brand kit. It immediately looked better, but since I'm going to be running paid ads, I needed to optimize the landing page to convert cold traffic. To do this, I went to mobbin.com to look for a landing page design that [music] I liked. Mobbin is a massive searchable database of user interface and user experience designs taken from the best apps in [music] the world. I took a screenshot, I put it into CoWork, and I asked it to analyze the landing page's layout and sales psychology elements, then [music] turn it into a PRD that I can apply to my app. And just like that, I have a clean, high-converting [music] landing page. I decided to position the app as astrology for business because this is the outcome men care about most. This also makes my TAM slightly bigger because I'm not explicitly excluding women from my messaging. The reason I'm making my landing page first is because it doesn't matter how good my product is if they never get past this page in the first [music] place. So, I spent most of day one iterating on it. Always prioritize market validation over the perfect product when building new software. Day two was all about product. I do need to have a usable product before I start running ads, so I created a low-friction onboarding. The promise of the product is to help them find the business that they were built for based off their chart. This is based on their zodiac sign, numerology, and Chinese [music] New Year animal. To calculate this, I added a form on sign-up to actually collect their free reading information, and then I connected the GPT 5.5 API to generate the reading. This outputs a free founder archetype and gives them a full reading on their personal chart. The goal of this is to build trust by giving them free value up front and to overcome any skepticism they may have about astrology by giving them a scary accurate reading. This reading should hit their ego and give them a big dopamine hit. So, this is where we introduce the paywall and ask for their credit card. To set up payments, I went to wop.sh. I copied this link and I installed the CLI into Claude. I then asked Claude to create two pricing options. One is $8.99 per week with a 3-day free trial, and the other is a $98 per year annual plan. Within 30 seconds, Claude created the products and embedded the checkout on my site. Now, I just need to design the payment modal and my onboarding is complete. To do this, I just asked Claude Co-work to go research the top astrology apps' payment flows, conversion tactics, and strategies, and then turn that into a PRD for my app. Co-work gave me this extremely premium-looking artifact that I now just download and give to Claude Code to implement. [music] So, now that I can charge my customers for the product, now I actually need to build the paid features. The front-end offer is to find the business that you are built for. Now, the paid offer is daily updates to help business owners make decisions based off their astrology and personality types. This is essentially a daily, weekly, and monthly astrology reading that my users can check every day to help them make business decisions. And I put it in this tab for users to see. Finally, to maximize the amount earned per user, I decided to add an [music] AI chat where users can ask questions and get hyper-personalized answers based off of their chart. So, now I had my free plan, my paid product, and my upsell ready to go. I just need to go to Vercel, buy a domain, and tell Claude to host my app. And in just two days, I have a fully functioning SaaS app live on the [music] internet, ready to get customers. Wednesday morning came around, and now it's time for the fun part. Marketing. I chose paid ads as my initial customer acquisition channel for two reasons. One, I can get data quickly and see what messaging the market responds to. And two, paid ads are the fastest way to get customers, and I'm doing a challenge, so that's important. To generate my ad creatives, I used the API and told Claude to look at my landing page and to generate five ad creatives [music] using the ad generator. Within a few minutes, I got the designs back, and oh my god, were they perfect. So, I had to generate me 15 more. That way, I had multiple to test with. Now it's time to actually run them. So I went to Wopp and decided to have the ads run on Meta because their AI targeting is top tier and I don't even need a Meta account to do this. I just hit this button, upload my logo and a banner, then enter my business name, and that's it. Wopp just spins up a Facebook page for me that I can now run [music] ads from. Now all I had to do is tell Claude to set up the actual ad campaign using the ad creatives we just created and my campaign was ready to go. To prove this works, I committed $500 of my own money to test with. I set the budget to $100 per day, I hit live, and my ads were already going out into the [music] world. I checked back in a few hours to see how they were performing and when I came back, at first I was shocked because I had already gotten hundreds of clicks, which is great, but then I looked a little deeper and saw the cost per click was only 3 cents, which may sound good, but that made me realize that I didn't have geo-targeting set. So Meta was likely sending my ads to a low-quality audience in third-world countries. As a result, I just lost $20 because I wasn't paying attention, which hurts because I should have my first customer by now, but in

days, but that required a full team, months of work, and a six-figure budget. So, for this challenge, [music] I wanted to see what would happen if I stripped it all away. No developers, no marketers, and no personal [music]

brand. Just me, AI, and one goal. Take a random app idea and turn it into a real software business by getting my first customer as fast as possible. To start, I needed to find a gap in the market to decide what

I should build. And so, I started where anyone in my position would start, and I started looking at what people are currently buying. I went to the App Store and saw that this astrology app had over 200,000 ratings. And after

looking into the astrology market, I realized that it's a $12.8 billion industry that's growing 20% year over year. As I scrolled the category, I noticed almost all of them target women, and they primarily focus on relationships. Literally none target

men. So, there's only two explanations for this. Either A, men don't pay for astrology and it's a dead market, or B, this gap in the market is my opportunity. And after doing a little bit more digging, it turns out that 18%

of men actually believe in astrology, which is about a billion people, including President Ronald Reagan and billionaire banker J.P. Morgan. This gave me the data I needed to feel confident in building the MVP as a web app. When I started building the app, I

used Claude Co-work [music] plus Wasp CLI and started describing all of the features that I wanted in the app. Then, I asked Co-work to turn this into a product requirement doc so I could take all of that information and turn it into

something that lists everything I want built. I then took that product requirement doc and paste it into Claude Code, and within just a few minutes, I had a functioning V1. But, it looked mega vibe-coded. So, I spent the next

few hours designing the app to make it feel [music] more premium. To do this, I first needed a brand guideline that would make the app feel more masculine.

So, I went back to Claude Co-work and asked it to build me a brand kit. I wanted a blend of astrology imagery with colors that implied wealth and in a few minutes, [music] I got a custom brand guidelines that defined brand font,

colors, and more. Then, I told Claude Code to redesign the V1 using the new brand kit. It immediately looked better, but since I'm going to be running paid ads, I needed to optimize the landing page to convert cold traffic. To do

this, I went to mobbin.com to look for a landing page design that [music] I liked. Mobbin is a massive searchable database of user interface and user experience designs taken from the best apps in [music] the world. I took a

screenshot, I put it into CoWork, and I asked it to analyze the landing page's layout and sales psychology elements, then [music] turn it into a PRD that I can apply to my app. And just like that, I have a clean, high-converting [music]

landing page. I decided to position the app as astrology for business because this is the outcome men care about most.

This also makes my TAM slightly bigger because I'm not explicitly excluding women from my messaging. The reason I'm making my landing page first is because it doesn't matter how good my product is if they never get past this page in the

first [music] place. So, I spent most of day one iterating on it. Always prioritize market validation over the perfect product when building a new software. Day two was all about product.

I do need to have a usable product before I start running ads, so I created a low-friction onboarding. The promise of the product is to help them find the business that they were built for based off their chart. This is based on their

zodiac sign, numerology, and Chinese [music] New Year animal. To calculate this, I added a form on sign-up to actually collect their free reading information, and then I connected the GPT 5.5 API to generate the reading.

This outputs a free founder archetype and gives them a full reading on their personal chart. The goal of this is to build trust by giving them free value up front and to overcome any skepticism they may have about astrology by giving

them a scary accurate reading. This reading should hit their ego and give them a big dopamine hit. So, this is where we introduce the paywall and ask for their credit card. To set up payments, I went to wop.sh. I copied

this link and I installed the CLI into Claude. I then asked Claude to create two pricing options. One is $8.99 per week with a 3-day free trial, and the other is a $98 per year annual plan.

Within 30 seconds, Claude created the products and embedded the checkout on my site. Now, I just need to design the payment modal and my onboarding is complete. To do this, I just asked Claude Co-work to go research the top

astrology apps payment flows, conversion tactics and strategies, and then turn that into a PRD for my app. Co-work gave me this extremely premium-looking artifact that I now just download and give to Claude Code to implement.

[music] So, now that I can charge my customers for the product, now I actually need to build the paid features. The front-end offer is to find the business that you are built for.

Now, the paid offer is daily updates to help business owners make decisions based off their astrology and personality types. This is essentially a daily, weekly, and monthly astrology reading that my users can check every day to help them make business

decisions. And I put it in this tab for users to see. Finally, to maximize the amount earned per user, I decided to add an [music] AI chat where users can ask questions and get hyper-personalized answers based off of their chart. So,

now I had my free plan, my paid product, and my upsell ready to go. I just need to go to Vercel, buy a domain, and tell Claude to host my app. And in just 2 days, I have a fully functioning SaaS

app live on the [music] internet, ready to get customers. Wednesday morning came around, and now it's time for the fun part. Marketing. I chose paid ads as my initial customer acquisition channel for two reasons. One, I can get data quickly

and see what messaging the market responds to. And two, paid ads are the fastest way to get customers, and I'm doing a challenge, so that's important.

To generate my ad creatives, I used the API and told Claude to look at my landing page and to generate five ad creatives [music] using the ad generator. Within a few minutes, I got the designs back, and oh

my god, were they perfect. So, I had to generate me 15 more. That way, I had multiple to test with. Now it's time to actually run them. So I went to Wopp and decided to have the ads run on Meta

because their AI targeting is top tier and I don't even need a Meta account to do this. I just hit this button, upload my logo and a banner, then enter my business name, and that's it. Wopp just spins up a Facebook page for me that I

can now run [music] ads from. Now all I had to do is tell Claude to set up the actual ad campaign using the ad creatives we just created and my campaign was ready to go. To prove this works, I committed $500 own

money to test with. I set the budget to $100 per day, I hit live, and my ads were already going out into the [music] world. I checked back in a few hours to see how they were performing and when I

came back, at first I was shocked because I had already gotten hundreds of clicks, which is great, but then I looked a little deeper and saw the cost per click was only 3 cents, which may sound good, but that made me realize

that I didn't have geo-targeting set. So Meta was likely sending my ads to a low-quality audience in third-world countries. As a result, I just lost $20 because I wasn't paying attention, which hurts because I should have my first

customer by now, but instead I'm in the red. I fixed my ads so they were only running in the four major English-speaking countries and turned them back on before going to sleep. On day four, I woke up really excited and

checked the ad account. This is the moment we'd all been waiting for. Did I just waste the last 3 days of my life or am I about to be on a yacht in Miami? I opened the account and there it was,

[music] my very first user to enter their credit card details. I got over 100 visitors, 32 people generated their charts, and one person signed up for the paid plan. Words can't describe how that made me feel because I remember on day

two, I started questioning everything. [music] In bed, I just kept having really negative thoughts. Was this just a stupid idea that only [music] I was excited about? Is there even any real market interest in this? And no way was

I the first person to try this. But I didn't listen to those thoughts and as a result, a validated business idea with a real user in just 4 days. Now, to level set here, I spent $40 on paid ads to get

a user on an $8.99 per week plan, which means I'm nowhere near profitable yet, but it's a good sign that A, there's market demand, and B, I created enough value that a random stranger on the internet was willing to pull out their

wallet and enter their credit card for my product, which is always a cool feeling. Now, considering I wasted my first $20 on low-quality traffic, and my conversion rate has a lot of room for improvement, this gave me the confidence

I needed to [music] keep pushing. Now, it was time for the real test. Can I become profitable? To do this, we need to analyze each step of the funnel and just see where we can make optimizations. I spun up a metrics

dashboard [music] that shows the funnel conversion at each step and saw an obvious weakness.

[music] The ads were performing great. Typically, a good click-through rate on an ad to cold traffic is around 3%. I had multiple creatives over 5%, and I even had one that was over 7%. So, I very clearly don't have a traffic issue,

but when I looked at the drop-off of people who started their chart and the ones who actually generated one, it was massive. I took a screenshot of these conversion rates [music] and asked Claude if this was normal. Claude gave

me some industry standard benchmarks at each step, and it turns out my suspicion was right. My onboarding completion rate was 20% lower than it should have been.

After digging a little deeper, I noticed that 75% of my traffic was on mobile web. So, I decided to test the product on my phone, and two problems became very clear. The location field in the form wasn't reading manual entries, and

when they input of their time on mobile, the field syntax required a colon, but it never communicated that. If these fields weren't filled out in the correct syntax, the button would stay unclickable. So, I asked Claude to fix

those two minor bugs, and the completion rate instantly started to rise, resulting in three more users entering their credit card information. So, at the end of day four, I had four users after spending $100 on ads, leaving us

at a $25 CAC, aka our cost for acquiring a customer. So, I gave Claude this information and asked what our CAC needs to be for us to be profitable based on our pricing model. And it came back with

under $20. It was getting quite late on day four, but I realized that my conversion rate from free reading to trial was only 1%. If we could just slightly boost this number, we could easily get our CAC down to where we need

it to be. So, I stayed up till 3:00 a.m. to set up email marketing. I don't expect every user to start their trial on their first session. So, to maximize my conversions, I'm sending each lead a daily custom reading directly to their

inbox every single morning to make sure they come back to the platform. Since I already collected their email in the onboarding, I just need to make a recent account and connect the Zapier MCP to set up my automated email flow. Then, I

just tell Claude to create the campaign. Now, since this is a daily custom reading for every user, this will add token costs, but it will result in a significant boost to my overall conversion rate, so it's well worth it.

Now, I'm pretty sure I fell asleep at my desk that night, but I woke up on day five super optimistic, only to check the ad account and see that I spent $100 on ads and got [music] no trials. Turns out there was a random

bug with my authorization and no one was getting paywalled, which hurt, but I'm not a quitter. So, I fixed the problem, made a few UI tweaks, and waited another day. I woke up on day six quite nervous because it's my last $100, but I checked

my WAB dashboard and it turns out [music] I got seven new trials in the last 24 hours. And if we do the math on $100 in ad spend with seven users, that leaves us at a $15 CAC, well below our

$20 goal for profitability. Now, since I'm doing a free trial model, we will need to see how many actually convert to paid, but I would expect at least 50% of these users to finish their trial, and I do expect the daily emails to drive a

few more conversions as the days go on. So, overall, I'm pretty happy with the results considering I had this idea less than a week ago, and the goal of the challenge was to see how fast I could go

from idea to first user and validate my idea. But, let me know your thoughts in the comment section below and if I should make a part two where I scale Monad to 10K MRR. Other than that, have a great day and I'll see you in [music]

the next one.