Ray (지범) Jang

Founder, CEO at Scenius 🚀 | ex-TikTok | Angel Investor

👋 Hi, I'm Ray Jang - Founder / CEO @ Atria (tryatria.com)

We help businesses find and make winning ads 10x faster using AI. 

If you want to access 500+ Canva templates to instantly make high-performing ads, 
sign up for our 5 day free trial.

Want to learn more? Let's talk:

ray@tryatria.com

---

Prior to starting Atria, my career was focused on building great products.

Highlights:

ByteDance / TikTok: Built a Rippling-like platform from 0 to 1;  increased the NPS score by 70%, saved ByteDance $6.4M USD in operational costs, and then commercialized to external enterprise customers. Promoted within 2 performance cycles to then work on TikTok Live.
Segment: Contracted to work on Segment's largest APAC enterprise account (Fortune 50). Helped to implement CDP so that CAC would be lowered by 28% and conversion rate would increase by 2.6x
OneDegree: Built the most profitable digital product line from 0 to 1
Flexport: Acted as the operational backbone for the fastest growing New York squad. Received 32/30 for outstanding performance.

Founder and CEO

tryatria.com

High-quality AI to boost ad conversions for e-commerce brands and ad agencies

Backed by Accel and Kleiner Perkins

A brand just spent $10M on an AI holiday ad. Pulled it in 72 hours after getting torn apart online.

Because you could 𝘧𝘦𝘦𝘭 it was lazy.

Here's what's happening right now that nobody's talking about:

"AI slop" mentions are up 9x in the last year. It's not a meme anymore, it's a movement.

Fast food chains pulling ads after backlash. Luxury brands getting called "soulless" in their own comment sections. People literally coined terms like "AI slop" to describe generic AI content.

The wildest part?

We spent billions building tools to 𝘢𝘮𝘱𝘭𝘪𝘧𝘺 human creativity...

And now "100% Human Made" is becoming a luxury label in marketing.

Think about that.

Here's what I'm seeing at Atria after running hundreds of creative tests:

The AI ads that perform aren't the ones that "look like AI."

They're the ones where AI handled the grunt work so humans could focus on the insight, the angle, the 𝘵𝘳𝘶𝘵𝘩 that makes people stop scrolling.

AI didn't kill creativity.

Lazy execution did.

𝗥𝗲𝗮𝗹 𝗾𝘂𝗲𝘀𝘁𝗶𝗼𝗻 𝗳𝗼𝗿 𝘆𝗼𝘂:

What's the worst "AI slop" ad you've seen this year and 𝘸𝘩𝘺 did it feel so off?

Drop it below. I'm collecting the best (worst?) examples for a breakdown next week.

Would love your take 👇

Follow me Ray Jang for daily posts as I scale Atria, an AI platform for marketers to create killer ads with ease.

Ray (지범) Jang 🐰

Making ad creatives simple with AI | Founder, CEO at Atria (tryatria.com) 🚀 | Forbes 30u30 | Ex-TikTok

USDA released their new food pyramid. Here's my ads pyramid.

The specifics of the best ad pyramid for you depends on your situation, but this hierarchy broadly applies to all brands.

𝗔𝘁 𝘁𝗵𝗲 𝗳𝗼𝘂𝗻𝗱𝗮𝘁𝗶𝗼𝗻: 𝗔 𝗖𝗹𝗲𝗮𝗿 𝗩𝗮𝗹𝘂𝗲 𝗢𝗳𝗳𝗲𝗿

Before fancy creatives, before viral hooks, before anything else: what are you actually selling and why should someone care? This is the handshake. The deal. The magnet that pulls people in. Without a compelling offer, everything else is just noise. Most brands skip this and wonder why their "creative" ads don't perform.

𝗧𝗵𝗲 𝗺𝗶𝗱𝗱𝗹𝗲 𝘁𝗶𝗲𝗿: 𝗘𝗺𝗼𝘁𝗶𝗼𝗻 + 𝗦𝗼𝗰𝗶𝗮𝗹 𝗣𝗿𝗼𝗼𝗳

Once you have a clear offer, you need two things to make people believe it:

Emotion and storytelling. The campfire moment. The relatable scenario that makes your product tangible. This is where people see themselves in your ad and think "that's me."

Data and social proof. The spreadsheet. The stars rating. The seal of approval. This is what makes skeptics believers. Real results, real customers, real numbers.

You need both. Emotion without proof is just hype. Proof without emotion is just a spec sheet.

𝗟𝗼𝘄𝗲𝗿 𝗺𝗶𝗱𝗱𝗹𝗲: 𝗧𝗵𝗲 𝗲𝘅𝗲𝗰𝘂𝘁𝗶𝗼𝗻 𝗯𝗮𝘀𝗶𝗰𝘀

The smartphone and spreadsheet represent mobile optimization and tracking. These are table stakes: vertical video, readable text, clean CTAs, measuring what matters. Important? Yes. But they only amplify what's already working above them.

𝗡𝗶𝗰𝗲 𝘁𝗼 𝗵𝗮𝘃𝗲: 𝗣𝗿𝗼𝗱𝘂𝗰𝘁𝗶𝗼𝗻 𝗽𝗼𝗹𝗶𝘀𝗵

The cinema camera, the sparkles, the sponge. These are extras.

A well-lit ad with a weak offer loses to an iPhone video with a strong one.

The industry has this pyramid upside down. We tend to optimize for what looks good in case studies instead of what moves the business forward.

Start at the top. Nail your offer. Add the proof and story. Then, and only then, make it pretty.

Follow me Ray Jang for daily posts as I scale Atria, an AI platform for marketers to create killer ads with ease.

OpenAI just dropped $250M on a brain-computer interface startup.

Here's what just happened:

OpenAI led the seed round for Merge Labs, valuing the company at $850M. The company is building non-invasive ultrasound tech that reads neural signals without surgery. No Neuralink-style brain implants.

The bet: as AI gets more capable, the way we communicate with it becomes the limiting factor.

OpenAI's official line was "Progress in interfaces enables progress in computing." They're saying the bottleneck isn't the AI anymore, it's how slowly we can get information in and out of it.

Right now, we type with our thumbs or talk out loud. That works fine for simple tasks. But if AI continues getting exponentially more powerful, that interface starts to feel like using a straw.

The vision here is collapsing the gap between thought and execution. You think of something and it happens. No typing, no clicking, no translating your idea into commands a computer understands.

First use case is medical (helping people with ALS or paralysis control computers). Long-term goal is making this available for everyone.

Will it work? No idea. BCI startups have a long history of big promises and limited results.

For context: imagine thinking of a creative concept and having it fully rendered without touching a keyboard. That's what they're building toward.

Follow me Ray Jang for daily posts as I scale Atria, an AI platform for marketers to create killer ads with ease.

Most performance marketers can tell you their best ad's ROAS down to the decimal.

But ask them "which combination of persona, hook, and angle drives the highest ROAS?" and you get assumptions and waffle.

Here's why:

You're stuck looking at performance one ad at a time. Ad A got 3.2 ROAS. Ad B got 2.1. Cool.

But you can't answer the specific questions that actually matter:

      •    Do "busy mom" personas convert better than "fitness enthusiast" personas?
      •    Does fear-based messaging outperform aspiration?
      •    Which visual hooks (product demo vs lifestyle shot) drive better results?

You have gut feelings. You have vague hunches. But you don't have data.

You're either optimizing blind or spending hours manually tagging hundreds of creatives in spreadsheets.

And most teams? They just skip it entirely and rely on "what feels right."

That's the gap between testing creatives and actually understanding what works.

We're about to close it.

Follow me Ray Jang for daily posts as I scale Atria, an AI platform for marketers to create killer ads with ease.

You're sitting on a goldmine.

And you don't even know it.

Most teams have run hundreds, sometimes thousands, of ads over the past year.

Every one of those ads is a data point. A signal about what resonates with your audience.

But here's the problem:

That data is completely unorganized.

You launched 50 ads in Q3. Some crushed. Some flopped. You moved on.

You tested new hooks in November. A few worked. You forgot why.

You ran holiday creative in December. It performed. You're not sure what to replicate.

This is creative debt.

You're accumulating tests but not extracting knowledge.

You're running experiments but not building a system.

Every ad you've ever launched contains insight about personas, hooks, emotions, angles, but it's buried in dashboards you'll never revisit.

So when it's time to plan Q1 creative, you're starting from scratch. Again.

You have gut feelings. Vague hunches. But no real data on what's actually worked across hundreds of tests.

Here's what changes with AI tagging:

Instead of creative debt, you build creative memory.

Atria's AI automatically tags every ad you've ever run by persona, hook, theme, emotion, visual style, offer type.

Then it shows you patterns.

Which personas drive the highest ROAS across all your tests. Which emotional angles convert best for different audience segments. Which creative themes have staying power vs. quick burnout.

You're not guessing anymore. You're learning.

Every test you run compounds. Every ad you launch adds to your creative intelligence.

The teams that have this figured out have a huge advantage.

They're not guessing. They know.

AI tagging is live in Analytics. It turns your creative history into clear insights.

Stop building debt. Start building memory.

Big update incoming.

Follow me Ray Jang for daily posts as I scale Atria, an AI platform for marketers to create killer ads with ease.

Anthropic just gave Claude folder access.

You can now point it at a folder on your Mac, and it will read, edit, create, and organize files without you touching anything.

This isn't "AI helps you write." This is "AI does the work while you sleep."

The feature is called Cowork, and it turns Claude from a chatbot into a junior employee.

𝗪𝗵𝗮𝘁 𝗰𝗵𝗮𝗻𝗴𝗲𝘀:

Instead of:

      •     Upload file → Get response → Copy output → Repeat

You do:

      •     Share folder → Define task → Claude handles it

Example: Share a folder of customer feedback emails and have it identify the top 5 feature requests with supporting quotes. Claude reads everything and outputs the report.

𝗧𝗵𝗲 𝗿𝗲𝗮𝗹 𝘁𝗲𝘀𝘁:

Will people actually trust it enough to let AI edit their files unsupervised?

Right now, most teams are comfortable asking AI for suggestions. Letting it take action in your workspace is a different level of trust.

But if this works, the bottleneck shifts.

The limit stops being execution speed. It becomes how clearly you can define what needs to happen.

Clarity becomes the most valuable skill. The AI handles the rest.

Follow me Ray Jang for daily posts as I scale Atria, an AI platform for marketers to create killer ads with ease.

𝗬𝗼𝘂 𝗵𝗮𝘃𝗲 𝗮𝗹𝗹 𝘁𝗵𝗲 𝗱𝗮𝘁𝗮.

𝗬𝗼𝘂 𝘀𝘁𝗶𝗹𝗹 𝗱𝗼𝗻'𝘁 𝗵𝗮𝘃𝗲 𝘁𝗵𝗲 𝗮𝗻𝘀𝘄𝗲𝗿𝘀.

Your ad account is sitting on mountains of performance signals. Audience behavior. Creative patterns. Spend efficiency trends.

But when you need to actually understand something - "Why did performance tank last Tuesday?" or "Which audience segment is bleeding budget?" - you're clicking through 47 tabs in Ads Manager hoping to stumble on it.

That's not analysis. That's археology with better dashboards.

𝗛𝗲𝗿𝗲'𝘀 𝘁𝗵𝗲 𝗴𝗮𝗽 𝗻𝗼𝗯𝗼𝗱𝘆'𝘀 𝘁𝗮𝗹𝗸𝗶𝗻𝗴 𝗮𝗯𝗼𝘂𝘁:

AI can generate your ads. Write your scripts. Edit your videos. Optimize your bids.

But it still can't tell you what's actually working in your account.

Strategy is still manual. Learning is still slow. Insights are still buried.

The teams that close this gap first are going to pull ahead fast.

𝗧𝗵𝗮𝘁'𝘀 𝘄𝗵𝘆 𝘄𝗲 𝗯𝘂𝗶𝗹𝘁 𝗥𝗮𝘆𝗮.

Raya is an AI analyst that lives inside Atria and knows every metric, every pattern, every anomaly in your ad data.

No filtering. No dashboard hunting. Just ask:

"Which audiences have the best ROAS this month?" "Why did my CPMs spike last week?" "What creative themes are outperforming?"

Raya answers in seconds.

Raya doesn't just surface data. It catches patterns you'd miss manually. It finds the insight buried three dashboards deep. It works 24/7 and already knows your entire account history.

Right now, Raya is live in Analytics and handles ad performance data. Over the next few weeks, we're expanding it across the platform.

𝗥𝗮𝘆𝗮 𝗶𝘀 𝗶𝗻 𝗯𝗲𝘁𝗮 𝗳𝗼𝗿 𝗮𝗹𝗹 𝗔𝘁𝗿𝗶𝗮 𝘂𝘀𝗲𝗿𝘀.

Go to Analytics. Start asking questions.

We're closing the AI gap in performance marketing. One question at a time.

Follow me Ray Jang for daily posts as I scale Atria, an AI platform for marketers to create killer ads with ease.

Meta changed how ads get shown. Most advertisers haven't noticed yet.

The update is called Andromeda. It rolled out quietly in late 2025, and agencies are only now confirming what's happening under the hood. Industry intelligence reports from AllMyClicks, Iternum Digital, and Clutch released between January 5-12 have detailed the technical breakdown.

𝗪𝗵𝗮𝘁 𝗰𝗵𝗮𝗻𝗴𝗲𝗱:

Meta used to prioritize who the user is (interests, demographics, behavior) when deciding which ads to show. Now it weighs what your ad looks like (visuals, copy, format) much more heavily.

The algorithm reads your creative with computer vision to predict engagement. Your targeting settings still matter, but creative quality has become the primary filter.

𝗪𝗵𝘆 𝗶𝘁 𝗺𝗮𝘁𝘁𝗲𝗿𝘀:

If your ad creative is weak, tightening your audience targeting won't save it. The system evaluates if your ad is worth showing based on creative signals before your audience settings even come into play.

Three things now kill your delivery in real time:

      1.   𝗧𝗵𝘂𝗺𝗯-𝘀𝘁𝗼𝗽 𝗿𝗮𝘁𝗲: If people scroll past in the first 3 seconds, Meta flags it as low signal and throttles reach immediately.
      2.   𝗥𝗲𝘁𝗲𝗻𝘁𝗶𝗼𝗻 𝗿𝗮𝘁𝗲: Watching 75% of your video matters more than clicking. Saves and shares outweigh passive clicks.
      3.   𝗠𝗲𝘀𝘀𝗮𝗴𝗲 𝗰𝗹𝗮𝗿𝗶𝘁𝘆: If your hook doesn't match your landing page, or if you're using fake curiosity to bait clicks, you get marked "below average" and pay a premium.

𝗧𝗵𝗲 𝗲𝘅𝗽𝗲𝗻𝘀𝗶𝘃𝗲 𝗽𝗮𝗿𝘁:

Ads with below-average quality rankings now pay 20-30% higher CPMs. And if your creative quality score drops low enough, Meta can exclude you from the auction entirely, even if you're willing to bid high.

You can't buy your way out of bad creative anymore.

𝗪𝗵𝗮𝘁 𝗯𝗿𝗲𝗮𝗸𝘀 𝗻𝗼𝘄:

Testing 10 versions of the same ad with small tweaks (different headlines, same video) doesn't work. Meta groups them under one "Entity ID" using visual DNA and treats them as a single test.

The system wants conceptual diversity. One UGC-style ad, one polished ad, one meme format. Not five variations of the same thing.

𝗪𝗵𝗮𝘁 𝘁𝗵𝗶𝘀 𝗺𝗲𝗮𝗻𝘀:

Interest targeting and demographic layering are getting weighted less in the algorithm. Meta is pushing broad targeting as the default because the system is now better at finding audiences by analyzing your creative than by following your manual settings.

If your ad looks like everyone else's, Meta has no reason to show it. The cost of being generic just went up.

The shift: stop obsessing over audience settings. Start making ads people actually want to watch.

Follow me Ray Jang for daily posts as I scale Atria, an AI platform for marketers to create killer ads with ease.

You know the process that's broken. You know the strategy work that keeps getting pushed to next quarter.

We skip the things we don't want to do.

And those are usually the exact things we need to do.

Same reason most people skip leg day.

Not because they don't know it matters. They skip it because it's hard, it's uncomfortable, and the results take forever to show up.

The hard stuff doesn't get easier by avoiding it. It just compounds.

So do the leg day.

In the gym. In your business. Wherever you've been skipping reps.

𝗧𝗵𝗲 𝗧𝗶𝗸𝗧𝗼𝗸 𝗱𝗲𝗮𝗹 𝗷𝘂𝘀𝘁 𝗰𝗹𝗼𝘀𝗲𝗱.

𝗔𝗻𝗱 𝗶𝘁 𝗰𝗵𝗮𝗻𝗴𝗲𝘀 𝗺𝗼𝗿𝗲 𝘁𝗵𝗮𝗻 𝘆𝗼𝘂 𝘁𝗵𝗶𝗻𝗸.

As of this morning, the Oracle-led consortium officially owns TikTok US. The 18-month legal saga is over. The ban threat is legally zero.

For most people, this is just "TikTok news."

For performance marketers, this is the starting gun on a CPM race.

𝗛𝗲𝗿𝗲'𝘀 𝘄𝗵𝗮𝘁 𝗷𝘂𝘀𝘁 𝗰𝗵𝗮𝗻𝗴𝗲𝗱:

For two years, Fortune 500 brands have been treating TikTok like a gamble. They flighted budgets month-to-month. They refused annual commit deals. They applied what insiders called the "uncertainty tax."

Why commit big dollars to a platform that might not exist in six months?

That hesitation kept institutional money on the sidelines. Finance brands. Insurance giants. Enterprise SaaS. The heavy spenders stayed in their safe harbors: Meta, Google, LinkedIn.

TikTok's auction was dominated by DTC and e-commerce brands willing to take the risk.

That discount is gone now.

𝗧𝗵𝗲 𝗙𝗲𝗯𝗿𝘂𝗮𝗿𝘆 𝗳𝗹𝗼𝗼𝗱 𝗶𝘀 𝗰𝗼𝗺𝗶𝗻𝗴.

TikTok is now designated as a "US Safe Harbor" platform. It has the Oracle badge. The compliance teams that blocked budgets for two years just gave the green light.

Institutional capital is about to enter the auction for the first time.

Brands that were legally barred from the platform can now compete. The "heavy money" that's been sitting on Meta is about to diversify.

And when institutional budgets hit an auction that's been running lean, CPMs don't stay flat.

𝗜𝗳 𝘆𝗼𝘂'𝘃𝗲 𝗯𝗲𝗲𝗻 𝗿𝗶𝗱𝗶𝗻𝗴 𝗰𝗵𝗲𝗮𝗽 𝗧𝗶𝗸𝗧𝗼𝗸 𝗮𝘁𝘁𝗲𝗻𝘁𝗶𝗼𝗻, 𝘁𝗵𝗮𝘁 𝘄𝗶𝗻𝗱𝗼𝘄 𝗶𝘀 𝗰𝗹𝗼𝘀𝗶𝗻𝗴.

The risk premium you were benefiting from is evaporating in real time.

Q1 is going to be very interesting.

If TikTok has been working for you, lock in annual commits now before the institutional wave hits. If you've been waiting on the sidelines, you just missed the discount window.

The platform didn't get banned. It got legitimized.

And legitimacy is expensive.

Follow me Ray Jang for daily posts as I scale Atria, an AI platform for marketers to create killer ads with ease.

One year ago this week, we launched Radar V2.

The pitch was simple: stop doing creative iterations manually. Let the AI handle the grunt work so you can focus on the big ideas.

We launched with:

      •    Unlimited creative iterations across all your client accounts
      •    60-day money-back guarantee if you don't see winners
      •    Competitor Spy Tool to reverse-engineer what's working in your niche
      •    Creative Reporting to track performance and spot emerging trends
      •    Review Mining to turn customer feedback into ad angles

A year later, we've shipped dozens of improvements, expanded capabilities, made it faster, smarter and more useful than ever. The success of this is echoed by the feedback we get from our customers.

The product pipeline for 2026 is stacked.

More control over your own data. Uncovering patterns about your product and audience you didn't even know existed. And a number things I can't even talk about yet, but will most certainly impact the entire ad landscape.

Your job is coming up with ideas that win. My job is making sure you have the tools to do it better.

Anyway, tryatria.com if you're into that sort of thing.

DeepSeek just quietly beat Google in Vision.

DeepSeek just open-sourced their OCR 2 model. While everyone was distracted by the "Chatbot Wars," they quietly commoditized computer vision.

Here's what happened yesterday (Jan 27):

DeepSeek released OCR 2 with a new architecture called "Visual Causal Flow." Instead of reading documents in a dumb, fixed grid pattern (left-to-right), it dynamically reorders visual tokens based on the document's logical structure: mimicking how the human eye naturally scans a page.

The result: It outperforms Google's Gemini 3 Pro on the OmniDocBench v1.5 benchmark (specifically on edit distance) and runs locally on a single consumer GPU.

Translation for businesses: The cost of extracting text from invoices, receipts, and messy creative assets just dropped to near zero. If you're still paying for AWS Textract or Google Vision API for basic document parsing, you're burning margin.

Follow me Ray Jang for daily posts as I scale Atria, an AI platform for marketers to create killer ads with ease.

taplio

Ray Jang's Linkedin Analytics

Ray Jang

Performance ScoreRay Jang's LinkedIn Performance

What is Ray talking about?

Ray Jang's Best Posts (last 30 days)

Ready to grow your LinkedIn presence?