Introducing Sett for Mobile Game UA
For years, I've watched the same thing happen across mobile gaming.
A studio builds a beautiful game. They run UA. Creatives perform. ROAS holds. Then it slips, CPIs/CPMs constantly rise. Then it slides. Then the team is in a code red or a war room arguing about why.
Then UA allocation for that title starts dropping and it never cracks its ceiling, or simply dies out.
Assuming the game is great from a product and monetization perspective, creative fatigue or lack of winner creatives is the diagnosis everyone reaches for. More creatives are needed, and more importantly winners are needed.
Moreover, few years ago with the rise of Applovin and the new generation of ML-based ad networks the problem became even worse.
Now, you have access to cutting-edge ad networks that are basically mass scale recommendation engines that can find high LTV players for your game. But they don't work unless you feed them with an ungodly amount of diverse, new concepts of playables and videos.
The charts tells a clearer story. A handful of studios consistently win at UA. The rest get washed. The difference comes down to a UA edge they hold, and a process that creates that edge.
The winners run systems. They test dozens of fresh concepts every cycle. They learn. They compound. They kill what doesn't work and let the networks pour fuel on what does. The losers run guesses. One playable a month. Six rounds of revisions. A senior creative's opinion as the tiebreaker.
We've known this for a decade. Nobody could do anything about it. Building dozens of playable concepts a month at production quality is not a thing a human team can do. Not at any reasonable cost. Not at any reasonable speed.
Today, that changes.
Data-driven playables at scale, made by AI agents
We're opening the waitlist for a system that operates the way winning UA programs operate, autonomously. Research the market. Generate concepts. Build them. Ship them. Measure them. Learn. Generate the next batch, smarter than the last.
One system, the entire loop.
Four things to understand about how it works.
1. It explores at scale.
97% of creatives fail. You can't beat that math with opinions. You beat it with volume.
The most successful mobile games don't get there by betting on the right concept before it ever met their audience in real life. They get there by running a system that tests dozens, finds the few that work, and double down on the winners.
AI agents that run the entire UA optimization loop, let you run that math without scaling headcount.
The system generates and builds at a pace humans cannot match. It doesn't sleep, argue, or get into ego fights.
2. It builds from scratch every time.
No template ceiling. No reskins.
Sett agents plan the mechanics, lay out the level design, then build each playable in real code. From your game assets, and with your permission true to your brand game art that's needed.
Every concept is a real fun game, built for that specific user motivation, that specific audience.
When you're exploring at volume, template ceilings kill you. Every concept hits the same wall. Building from scratch removes the wall.
3. It learns from every play.
Every playable ships with analytics baked in. Drop-off points. Engagement patterns. The moment a player gets bored. The moment they get hooked.
That data doesn't sit in a dashboard you'll never open. It feeds the system. The next concept is informed by every concept that ran before it. Compounding is the real unlock.
Every concept is smarter than the last.
More over Sett agents create, manage, and update knowledge and memory of what experiments you ran, how much share of spend concepts got over time, and your own team preferences and strategy, allowing UA teams to steer them towards greatness.
4. It frees humans to do what only humans do.
This is the part most people get wrong about AI in creative.
Agentic playables don't replace taste. They replace the parts of the job humans were never good at. Building dozens of playables a month. Holding the memory of what worked six tests ago. Running mechanical labor at machine speed.
Humans have always been great at the creative leap. At knowing when something feels right. At setting direction and making the hard judgment calls. The system handles the volume, the building, the measuring, the memory. You steer and handle the spark, and the strategy.
The combination is the point and what truly brings AI agents from a lab, to at-scale production ready system working along your UA efforts.
The waitlist opens today
If you run UA, or deal with UA creatives for a mobile game, this is built for you.
Book a demo → · Product overview →
See it live, learn how Sett can build this machine for your studio, and watch spend scales.
About the Author
Fishi is the Head of Marketing at Sett. His brain is a chaotic jukebox of ideas with more cultural references than any feed can handle. He collects sneakers and plays chess while you’re still counting sheep.
