Ray Jang found product-market fit on a mountain top
with Ray Jang, Atria
Ray Jang found product-market fit on a mountain top
Show Notes
Ray Jang is the founder and CEO of Atria, an AI-powered creative intelligence platform that helps brands and agencies optimize, generate, and launch higher-performing ads on Meta, TikTok, and YouTube. Ray came up as a senior product manager at TikTok, where he spent close to two years navigating one of the most competitive product cultures in tech before leaving to start a company.
His first attempt was a gaming platform - an ambitious alternative to Steam - that flatlined after roughly a year and a half. With three months of runway left and no clear plan B, Ray isolated himself completely, summited one of the highest peaks in Southeast Asia, and had the epiphany that became Atria: stop solving for his own growth problem and start building a platform that solves it for everyone else. The first version shipped in January 2024. Within the first week, it was making money.
The Mountain and the Pivot
The pivot story is worth unpacking in full because it illustrates how Ray actually thinks. When he decided to kill the gaming company, he did not immediately swap in a new idea. He went into what he describes as a cave. He cut caffeine, ate once a day, dropped social media, stopped going to parties, and lost friends in the process. He studied more than a thousand ideas a day across Crunchbase and Y Combinator - not to find the next thing to chase, but to develop enough conviction and contrast that the right direction would become obvious.
The key realization came on the summit of Mount Canlaon in Southeast Asia: the most painful thing about building the gaming company had been figuring out distribution. Why not build the platform that solves that pain for everyone? The insight was personal, the timing was desperate, and the product emerged from the intersection of both. That is the pattern Ray now teaches - find a problem you can feel, that a customer will pay for on day one, with visible time-to-value and no regulatory landmines blocking the path.
What Atria Actually Does
Atria started as a competitor intelligence tool - a way to spy on what ads rival companies were running. It has since evolved into a full creative intelligence stack covering three phases: optimize, create, and launch. The launch layer (auto-deploy to Meta and TikTok, with YouTube coming in Q1) is the newest addition and closes the loop from insight to live ad without leaving the platform.
The core product is Radar - an in-house scoring algorithm that analyzes video ad creatives using computer vision. Radar breaks down each ad into its building blocks (hook strength, retention curve, click-through rate signals, conversion indicators), assigns a letter grade to each component, and delivers tailored recommendations for improvement. For a brand running thousands of ads simultaneously, Radar answers the question no human team can answer at scale: which of these is actually worth doubling down on right now?
Clone Ads is the creation layer. A user browses the Meta Ad Library, finds an ad style that resonates, clicks clone, pastes their website URL, and Atria generates a creative brief automatically - then produces ads in the brand's tone. The goal is not to generate AI slop. Meta is already penalizing undifferentiated AI content. The goal is to inject more strategic thinking and human-inspired ingenuity behind each creative, so marketers can do more with less.
The Darwinian Mutation Logic for Ad Creativity
One of Ray's most operationally useful ideas is how Atria avoids the trap of recycling a brand's own stale ad patterns. His framing borrows from evolutionary science: in any mutation, you have to introduce a controlled amount of randomness grounded in principles the brand has not yet tried. Simply copying what a direct competitor does caps out quickly and pushes the brand toward “same as same” in the eyes of its audience.
The better move is cross-industry inspiration. A B2B SaaS project management tool should not only study Monday.com - it should study Grammarly, Sprout Social, or a completely unrelated category with strong direct response performance. Atria facilitates this by connecting a brand's historical ads manager data (what they've already tried) with a broader library of non-adjacent creative examples, so the recommendations are both novel and grounded in what actually converts.
AI as a Binary Filter for Craft
Ray's take on AI and jobs is sharper than the usual framing. He argues that AI has not simply automated tasks - it has created a binary filter on whether you actually belong in your craft. Before AI, there was enough slack in the system that someone could fluff their way through a creative or marketing role. There was no strong alternative pressing on their output. AI removed that slack.
The practical consequence runs in two directions. For people who were just paying the bills without deep investment in the craft, AI surfaces that misalignment faster than any performance review. For people who truly care about the work, AI reduces the gatekeeping - no more mandatory dues-paying at Ogilvy or WPP before you can earn a seat. A 22-year-old with genuine creative instincts can now prove it without a resume. The floor drops (more people can enter), the ceiling rises (the best can reach further), and the bloated middle hollows out. Ray considers this net positive.
From Top-Down Conviction to Bottoms-Up Learning
The most honest thing Ray says about his first two years of company building is that he was too top-down. He had a thesis, and the rest of the world was supposed to conform to it. The pain of watching that not happen - and the discipline of the pivot process - rewired how he builds. He now operates on 90-day planning cycles rather than three-to-five year roadmaps, because the pace of concurrent accelerating trends (LLMs, computer vision, robotics, new form factors) makes longer-term determinism a liability rather than a strength.
The future visions he holds for Atria are held loosely: a PMF engine that can validate new SKUs with genuine buyer intent faster than customer calls; a proprietary advertising channel built on emerging form factors like AI chat interfaces and smart glasses; and the most insightful customer intelligence layer in the market, synthesizing census data, credit card signals, and CRM data into actionable persona maps. Which of these becomes the company's primary direction depends on what the data says in the next 90 days.
Tools & Resources Mentioned
- Atria - Ray's creative intelligence platform; tryatria.com (aiming to move to atria.ai)
- Radar - Atria's in-house scoring algorithm for video ad creatives; grades hook, retention, CTR, and conversion components
- Clone Ads - Atria tool for generating creative briefs and brand-matched ads from Meta Ad Library inspiration
- WisprFlow - voice-to-text tool Ray uses to draft emails and strategy docs by rambling, then cleans up in GPT or Claude
- Claude / ChatGPT Voice Mode - Ray uses voice mode during travel as an always-patient, ever-curious conversation and research partner
- Make.com / n8n - automation platforms Ray plans to explore for building out agentic workflows
- Krisp.ai - AI background noise cancellation + meeting notes tool mentioned by the host
- Meta Ad Library - primary inspiration source for Atria's Clone Ads workflow
- VaynerCommerce - Gary Vee's agency; cited as a current Atria agency customer
- True Classic - e-commerce brand spending 7 figures/month on Meta; cited as a key Atria brand customer win
Frameworks
The PMF Science Filter
Ray's framework for evaluating new ideas after his first company failed: go after a large market, target a customer you can directly address, avoid heavy regulation, aim for high ACV, ensure day-one payment is possible, require clear time-to-value, make sure AI workflow effects are visible, and build for a pain you personally feel so you can sustain the journey. Any idea that fails one of these filters gets cut.
AI as Binary Filter on Craft
AI has eliminated the middle ground in creative and knowledge work. Workers who were coasting without deep craft investment are now exposed - the alternative is too strong. Workers who genuinely treat their field as a craft find the floor lowered (easier to enter) and the ceiling raised (further reach). The middle hollows out. This is Ray's framework for understanding AI's impact on labor markets without defaulting to either doom or hype.
Darwinian Mutation for Ad Creative
To avoid recycling stale brand patterns, introduce controlled randomness sourced from ad principles and angles the brand has not yet tried. Ground the randomness in the brand's historical data (what they've already run) and draw inspiration from non-adjacent industries rather than direct competitors. Copying only within your category caps out quickly and makes you same-as-same in the audience's perception.
Bottoms-Up Learning Over Top-Down Conviction
Ray's lesson from his first company: conviction is necessary but not sufficient. The world has its own data and reality. Good company building merges founder vision with real-world signals. Operating on 90-day cycles instead of multi-year plans forces regular recalibration of assumptions rather than pushing reality to conform to a fixed thesis.
Systems Thinking as the Core AI Leverage Point
The founders who unlock the most value from AI are those who can think in systems - identifying processes, mapping if/then logic, and designing workflows that AI workers can execute. Automation and no-code tools are a useful training ground because they force you to externalize and codify your own thinking before AI can act on it.
FAQ
What is Atria and who is it built for?
Atria is a creative intelligence platform for brands and agencies running paid ads on Meta, TikTok, and YouTube. It helps with all three phases of the creative workflow: optimizing existing ads (Radar scoring), generating new ads (Clone Ads + creative brief generation), and launching them directly from the platform. Core customers include direct-to-consumer e-commerce brands, B2B SaaS companies, and agencies managing multiple client ad accounts.
What is Radar and how does it work?
Radar is Atria's proprietary scoring algorithm for video ad creatives. It uses computer vision to analyze a video frame by frame, breaking the ad into its functional building blocks: hook strength, retention curve, click-through rate potential, and conversion indicators. Each component receives a letter grade and a specific set of tailored recommendations. For large brands running thousands of simultaneous creatives, Radar answers the question no human team can answer at scale: which of these is worth doubling down on right now?
How did Ray go from a failing gaming company to Atria in three months?
With three months of runway left and a flatlining growth chart, Ray killed the gaming company, entered a period of complete isolation (one meal a day, no social media, no caffeine, no social events), and studied thousands of ideas on Crunchbase and Y Combinator until he had the conviction to move. The epiphany came while summiting Mount Canlaon: the hardest thing about the first company had been distribution - so why not build the platform that solves distribution for everyone? The first version of Atria shipped January 2024 and was profitable within the first week.
Why does Atria draw inspiration from non-adjacent industries rather than direct competitors?
Copying only within your own category caps out quickly. If every e-commerce brand studies only other e-commerce brands, the entire category converges toward the same visual language and hook formulas. The audience habituates and stops responding. Non-adjacent inspiration - a project management tool studying a skincare brand's creative, for example - introduces angles the audience hasn't seen in that context. Atria facilitates this by indexing a brand's historical ad data against a broad library of non-adjacent industry examples, sourcing novelty from contrast rather than from copying.
What does Ray mean when he says AI is a binary filter on craft?
Before AI, there was enough slack in creative and knowledge work that a person could do an adequate job without deep investment in the craft - there was no strong alternative to press on their output. AI removed that slack. Now the question is binary: are you genuinely treating this as a craft, or are you there to pay the bills and fluff? People in the second category are exposed faster than ever. People in the first category gain access they never had before - AI lowers the barrier to entry and raises the ceiling on what they can produce.
What is Ray's long-term vision for Atria?
Ray holds the future loosely and operates on 90-day cycles. The flavors he is exploring: a PMF engine that can validate new SKUs or product concepts with genuine buyer intent faster than customer calls; a proprietary advertising channel built on emerging form factors like AI chat interfaces and smart glasses; and a deep customer intelligence layer that synthesizes census data, credit card signals, and CRM data into actionable persona maps. He is explicit that any of these could change based on what the next 90 days of data shows.
What is Ray's favorite AI productivity tool?
WisprFlow - a voice-to-text tool that lets him speak his thoughts freely, then paste the transcript into GPT or Claude for cleanup into a structured email, product doc, or strategy note. Ray describes himself as a verbal thinker who used to face significant inertia in writing. WisprFlow removes the cold start problem: instead of staring at a blank page, he just rambles, and the cleanup is Claude's job.