12:08Deniz Sancar · Tutorial
I Asked Claude Code + GPT Image 2 To Make me Viral Ads
A 12-minute agency playbook for going from zero brief to ten finished static ads in under an hour using Claude Code and GPT Image 2.
May 23rdThe argument in one line.
In 2026, a $10,000/month AI SaaS is decided less by coding skill and more by three filters: a pain point that is financial and repetitive, a single narrow customer to build for, and a real distribution strategy.
Read if. Skip if.
READ IF YOU ARE…
- You're a non-technical or lightly technical builder considering an AI SaaS and want a filter for whether an idea is worth building.
- You're bootstrapped (not VC-funded) and keep second-guessing whether to niche down or stay broad to maximize addressable market.
- You've already shipped a SaaS with signups but few paying, engaged customers and aren't sure what to fix first.
- You want a concrete way to validate a pain point with numbers instead of gut feel before writing any code.
SKIP IF…
- You're VC-funded and already have the budget to go mass-market from day one.
- You're looking for a coding tutorial. This is a positioning and go-to-market framework, not a build walkthrough.
The full version, fast.
The creator behind Verlo (100,000 signups, six figures in 100 days) breaks down the exact four steps he'd repeat from zero. First, only build for a pain point that's financially or emotionally costly AND repeats often enough to notice, quantified in real numbers rather than vague complaints. Second, resist going mass-market: pick one narrow ideal customer and one specific problem so the copy, landing page, and product all read as built for that one person. Third, treat marketing as the real differentiator, tracking what's already working in your niche instead of guessing, and using volume outreach to creators for content. Fourth, once you have paying customers, prioritize ten engaged, paying users over a hundred free signups, because paid users tell the truth and free users have no reason to.
Free for members
Chat with this breakdown — free.
Sign in and you get 23 free chat messages on us — ask for the hook, quote a framework, find the exact transcript moment, generate a markdown action plan. Bring your own key when you want unlimited.
Create a free account →Where the time goes.
00:00 – 01:12
01 · Cold open: why distribution beats code
Establishes credibility (Verlo's 100k signups in 100 days) and states the thesis: a $10k/month AI SaaS in 2026 comes down to picking the right pain point, building for one person's problem, and marketing it properly. Coding ability is explicitly ruled out as the deciding factor.
01:12 – 03:28
02 · Step 1: Find a pain point that's financial and repetitive
A sellable problem needs two things: it hurts (financially, emotionally, physically) and it repeats often enough to notice. Walks through an agency losing $15k-$30k a week on slow lead scheduling, and names Calendly as the company that solved that exact pain with a single-purpose scheduling tool.
03:28 – 05:27
03 · Step 2: Build for one ICP, not everyone
Going mass-market requires a budget bootstrapped founders don't have; marketing to everyone becomes marketing to no one. Instructs narrowing a broad category (content-feedback tool for 'creators') down to one specific segment, like TikTok shop agencies, so the landing page and product read as built for that exact person.
05:27 – 07:41
04 · Step 3: Win on marketing and distribution
Distribution is framed as the real differentiator between an idea and a profitable business, illustrated with an AI football app that hit #1 free app in the UK App Store within days. Introduces Verlo (his own social listening tool for tracking viral TikTok/Instagram/YouTube Shorts trends) and Mockingbird (a tool for remixing top-performing Meta ads) as the distribution stack.
07:41 – 08:58
05 · Step 4: What to do with your first customers
Argues for prioritizing ten engaged, paying customers over a hundred disengaged free users, because paid users tell the truth and free users have no skin in the game. Includes a mentor's line that free advice is worth exactly what you paid for it: nothing.
08:58 – 09:26
06 · Recap and next video
Recaps the four steps in order and points viewers to a longer masterclass video on building AI SaaS in 2026.
Lines worth screenshotting.
- A problem only qualifies as sellable when it hurts financially, emotionally, or physically AND happens repeatedly enough for the person to notice.
- Quantifying a pain point in real numbers, like $3,000 to $5,000 per lost lead, makes it investable in a way 'it's a bit annoying' never will.
- A business losing seven $3,000-$5,000 leads a week to slow scheduling can be bleeding $15,000 to $30,000 every single week.
- Bootstrapped founders can't out-market VC-funded competitors by going broad, so mass-market positioning without a mass-market budget quietly becomes marketing to no one.
- Niching down doesn't close doors, it opens them. The riches are in the niches, not in trying to serve every industry at once.
- Visitors convert when the copy and reviews on a landing page make them think 'this was built for someone exactly like me,' which only happens with one narrow ICP.
- Ten paying customers who actually talk to you are worth more early-stage signal than a hundred free users who never engage.
- Free users have no skin in the game, so their product feedback can't be trusted the way a paying customer's can.
- Distribution, not coding skill, decides whether a good idea turns into a profitable SaaS.
- An AI-built football app hit #1 free app on the UK App Store within days, ranking above Fortnite and Roblox, purely on distribution strategy.
- Contacting hundreds of creators for user-generated content works on volume alone: even a low reply rate produces usable footage.
- Free trials lower signup friction, but giving a product away permanently to get feedback trains you on advice from people with nothing at stake.
The four-step filter for a $10k/month SaaS
THE 4-STEP FILTER
A sellable AI SaaS idea passes one filter (a pain that's financial and repeats), gets built for one narrow customer instead of everyone, and only becomes a business once distribution and paying-customer feedback take over.
02Step 1: Find a pain point that's financial and repetitive
- A problem only qualifies as sellable if it hurts financially, emotionally, or physically AND happens repeatedly enough for the person to notice it.
- Quantify the pain in real numbers, like $3,000-$5,000 per lost lead or five hours a week wasted, instead of vague complaints like 'it's annoying.'
- Research pain points on Reddit, X, and niche communities by asking questions, not pitching, before you write a line of code.
- Even in a crowded market you can win by being cheaper or faster than an existing solution, not just by finding an unsolved problem.
03Step 2: Build for one ICP, not everyone
- Going after every industry at once requires a marketing budget bootstrapped founders don't have, so mass-market positioning quietly becomes marketing to no one.
- Pick one narrow ICP and one specific problem before building anything; that single decision simplifies the landing page, copy, and product scope.
- A tool built for a broad category should still launch narrowed to its single highest-potential sub-segment, like TikTok shop agencies instead of all content creators.
- Visitors convert when the copy and reviews make them think 'this was built for someone exactly like me,' which only happens with a narrow ICP.
04Step 3: Win on marketing and distribution
- Distribution decides whether a good idea becomes a profitable business, not how good the underlying idea or code is.
- Before writing any ad or post, find which content themes are working in your niche right now rather than reusing tactics from six months ago.
- Contacting hundreds of creators for user-generated content works on volume alone; even a low reply rate nets usable footage.
05Step 4: What to do with your first customers
- Ten paying customers who actually engage with you are worth more early-stage signal than a hundred free signups who never talk to you.
- Paid users tell the truth because they have money on the line; free users have no skin in the game, so their feedback isn't trustworthy.
- Free trials are fine for lowering signup friction, but giving the product away permanently for feedback teaches you nothing reliable about what to build next.
Terms worth knowing.
- ICP
- Ideal Customer Profile: the single, specific type of customer a product is built and marketed for, as opposed to a broad, undefined audience.
- MVP
- Minimum Viable Product: the smallest version of a product that solves one problem for one customer type well enough to sell and learn from.
- Social listening platform
- A tool that tracks and analyzes public social media content (posts, trends, competitors) at scale to surface what's currently performing well.
Things they pointed at.
Lines you could clip.
01:02
“My friends, distribution is everything.”
03:46
“Now you could be losing anywhere from $15,000 to over $30,000 per week on these unconverted leads.”
04:10
“The riches are in the niches, and it's absolutely true.”
08:25
“I had a mentor once tell me that free advice is worth exactly what you paid for it. Nothing.”
Word for word.
Read-along
Don't just watch it. Burn it in.
See every word as it's spoken — crank it to 2× and still catch all of it. The same dual-channel trick behind Amazon's Kindle + Audible.
metaphoranalogystory
Right, so a couple years ago, my co -founder Nick and I built a software company called Verlo. Zero users, zero revenue, and to be honest, zero clue what we were doing half the time. Fast forward a couple years and we've crossed 100 ,000 signups and over six figures in revenue in the first 100 days.
And that was through organic marketing. Since then, I've also launched multiple other AI startups, including apps and softwares that collectively make me over six figures in profit every single year. And that's not to flex genuinely.
Because I'm going to show you in this video exactly what I would do if I had to start all over again building a software in 2026 using AI. Like step by step what I would do from scratch without any of my reputation behind me. Because here's the thing, in 2026, building a minimum $10 ,000 per month AI SaaS is not about how good of a coder you are.
To be honest, I can't code. I never could, and I probably never will. What it actually comes down to is three things.
Picking the right idea, building for one person's problem rather than everybody's problem, and then marketing it properly using the right tools in 2026. Because my friends, distribution is everything. I'm not talking about the stuff that worked back in 2022 pre -chat GPT.
These are 2020. six proven methods so get your notebook and get your pen let's get into this step by step Okay, step one.
And this is where most people mess up before they've even opened their computer. You do not need to create a new idea. You don't need the next Microsoft or the next Apple.
You just need a pain plus repetition. A pain point means it actually hurts financially, emotionally, mentally, physically, whatever it may be. And repetition means it happens over and over again.
That could be daily, weekly, monthly, enough for the person to notice. And if a problem ticks both those boxes, it's sellable. Simple as that.
Okay, so let's think about this from a business standpoint. Imagine an agency that is struggling to get leads on a call. That's not a one -time pain.
They need those leads to get onto an introductory call so they can sell them on their offer and get new clients. That is something that can happen multiple times a week, especially for a growing agency. They're playing tennis over emails trying to find the right time and by then the lead slips through the cracks and it's another unsigned client.
Now let's do some napkin math here. If each lead is worth $3 ,000 to $5 ,000 and on average you're losing out on seven leads per... week because you haven't got the right infrastructure in place to take them from a lead onto an introductory call.
Now you could be losing anywhere from $15 ,000 to over $30 ,000 per week on these unconverted leads. And that's a pain that is financially significant and repeated constantly. And what company solved this exact pain?
canonly canonly is not complicated whatsoever it fixes one problem a meeting scheduler but it does it in a way that is perfect for agencies and busy entrepreneurs now rather than saying what time works for you 5pm on a wednesday you just say here's my calendar booking a time that suits you so here's your homework and genuinely pause this video and think for a couple minutes go to reddit go to x go to school communities in your niche don't sell anything just ask people and look what are some problems people are facing because you'll be surprised that even in 2026 where it feels like every single startup has been created already, there are still genuine pain points that haven't been solved yet.
And even if a pain point has been solved, you can always innovate to be cheaper or quicker than others on the market and still steal some market share. And the most important thing is that when you're trying to find problems, don't look for qualitative or wishy -washy answers. Not stuff like, yeah, it's a bit annoying.
I'm talking about actual numbers. I'm losing out five hours per week on this, or it's costing me $500 per month on this. Those are real pain points you can build a product around.
And if you can quantify those problems into numbers they're far more sellable okay step two and this is where people go wrong after finding their pain point they try to build for everyone they go mass market super early now going mass market can work if you are a funded company with lots of vc backing but if you are bootstrapped like the vast majority of people watching this and like all of my companies you can't just go for everyone because unfortunately you'll need a humongous budget to market to everybody and if you market to everyone in reality you market to no one so don't be saying you go to freelancers agencies, small business owners, large conglomerates, You don't need all of that.
And worse, don't be saying you go to every single industry under the sun. It doesn't make you the attractive option. Niching down does not close doors.
In fact, it opens them. You may have heard the saying that and it's absolutely true. And if you want to build a 10K per month and more SaaS, you can easily do that by focusing on one specific target area.
You want deep and narrow, not wide and shallow. So let's say you're building an AI tool that gives feedback on content someone is posting. Now that's the type of tool that could apply to loads of people.
It could be Instagram creators, it could be TikTok creators, YouTubers, podcasts, you name it. But start by really honing in on one target audience, the one target audience that's got the biggest potential. So for example, you could create an AI tool that analyzes TikTok shop videos for TikTok shop agencies to show them exactly where the content is losing out and what they need to do instead to get sales.
So you've really narrowed in on the TikTok shop angle and specifically agencies around the TikTok shop because it's an unsaturated market for your tool to serve. And yes, you'll be surprised how... people from other niches also find the tool and find it to be valuable but that's not really your target at the start one icp solving one problem be dead specific that is your mvp and the reason for this is because once you've got that problem and that icp everything else becomes super easy from the landing page to the website copy to the product you actually build it all becomes hyper focused on that one problem because the goal is when the target audience lands on your website they see the copy the h1 they see the user reviews they think damn this is made for someone just like me and that is how you capture your target audience when they look at a tool and go that is built for them so start with what result does this person need how can we fix one problem that's stopping them get to that result and how can we do it faster or cheaper than others on the market okay step three marketing and this is what separates people that have incredible ideas from actual profitable businesses because in 2026 distribution is everything when i launched my own football game 38 oh no utilizing ai it got to 1 million users in the first five days and became the number one
free downloaded game on the App Store in the UK. It quite literally ranked above games like Fortnite and Roblox a whole weekend within the United Kingdom. And that's absolutely crazy to say.
Even mainstream media outlets were writing about the success of my game. And that was largely down to my distribution strategy. And I'm going to share some of the points I used with you.
Now, firstly, before you write a single ad or make a single post, you need to know exactly what sort of content themes are working in your niche right now. I'm not talking about the strategies that worked six months ago. I'm talking about up to day, viral content trends that you can leverage when it comes to the posts that you create.
And that is where I'm going to let you into a little secret. That is the exact reason I built Verlo. You see, Verlo is a social listening platform that tracks millions of videos across TikTok, Instagram, and YouTube Shorts to show you exactly what content themes are working best.
You create a content research agent whose only job is to go away and find you viral content, and then you can leverage those content themes to create your own hooks, scripts, and content calendars. And Verlo goes a step further because of all the data points analyzed on the platform. you can ask vcopilot some questions like i'm targeting footballers between the ages of 20 and 30 can you please find example content for these sort of people and vcopilot will go away and do exactly that literally find example ugc for your industry and then what you do that data is totally up to you you can record your own videos or even contact those creators to promote your brand if you contact hundreds of creators at least one will say yes and if you want to make some image ads for meta to go along with this i'd recommend checking out mockingbird which showcases the best performing meta ads on the meta ad library having helped you recreate them with one click for your particular brand i'll link both verlo and mockingbird down below in the description and in the pinned comment both come for seven day free trial so you can test out all of this completely risk -free without paying a
Okay, step four, what to do with your first customers. Now, here's the part almost nobody really wants to hear. But to be honest, it's probably the most important section of this video to go from $100 per month to over $10 ,000 per month.
And that is that your first 10 paying customers. are nothing to do with the money i'd rather you make a grand in your first 90 days with 10 engaged paying customers then make 5k with 100 customers who don't even talk to you once paid users tell you the truth and if they're genuinely engaging with you as a founder that means they're interested in the product and want it to help them and another thing never take advice from free users free users lie to you and it's not malicious it's just they have no skin in the game because they've put no money down if someone's not willing to spend 10 20 50 on your product then what good is their advice I had a mentor once tell me that free advice is worth exactly what you paid for it.
nothing that's why i recommend never give away your product for free to get feedback unless it's some kind of mentor or someone's opinion you genuinely value feel free to do a free trial like i do some of my products as it's a good way of getting people in to test the product first before committing to signing up but you should never ever ever take that free advice seriously so to bring it all together step one is to find a pain point that is genuinely repetitive then step two build for one icp with that one exact problem niching down step three be aggressive with marketing and distribution and you can Utilize Volo to help you see exactly what is trending for your particular niche.
And then step four, obsess over feedback with those first 10 paying customers and use it to make your product perfect. That isn't theory. That's exactly what I've done with the softwares I built.
And if you follow that step -by -step plan, you are well on your way to making your first profitable software. And if you want to go a step further and see a masterclass that I delivered on how to build AI SaaS in 2026, you've got to check out the video right beside me. Might be the best one hour you'll spend this week.
I'll see you there.
The Hook
The bait, then the rug-pull.
He grew one SaaS from zero users to over 100,000 signups and six figures in revenue inside 100 days, using organic marketing alone. Here he rebuilds the exact four-step process from scratch, minus the reputation and the head start.
Named ideas worth stealing.
01:12concept
Pain + Repetition filter
- Pain: hurts financially, emotionally, mentally, or physically
- Repetition: happens often enough (daily, weekly, monthly) for the person to notice
A two-part test for whether a problem is sellable: it must genuinely hurt AND recur, not just be a one-time annoyance.
Steal forvetting any new product idea before writing a line of code
04:58concept
One ICP, one problem
Narrow the target customer and the problem solved down to one specific pairing before building anything; that single decision simplifies the landing page, copy, and product scope.
Steal forpositioning an MVP so it converts on first visit
00:00list
Four-step SaaS founding system
- Find a pain point that's financial and repetitive
- Build for one ICP, not everyone
- Win with distribution and marketing
- Obsess over feedback from your first ten paying customers
The full framework the video walks through step by step.
Steal forany early-stage SaaS go-to-market plan
How they asked for the click.
VERBAL ASK
09:05next-video
“if you want to go a step further and see a masterclass that I delivered on how to build AI SaaS in 2026, you've got to check out the video right beside me”
Soft cross-promotion after the recap, framed as a natural next step rather than a hard sell.
MENTIONED ON CAMERA
FROM THE DESCRIPTION
PRIMARY CTAWhere the creator wants you to go next.
AFFILIATECommission earned if you click.
Visual structure at a glance.
