The character was the most consequential early decision. The 3D asset the product team had developed didn't read well for a casual audience — it lacked the expressiveness and warmth the game needed. Rather than wait for product to resolve it, I directed the team to rebuild the character direction ourselves. We used Nano-based prompting workflows to rapidly develop and evaluate concepts, then brought the chosen direction to life through Kling AI animation. This was a deliberate choice to treat AI-assisted workflows as a legitimate concepting and production method — not a workaround, but a considered approach suited to the constraints we were operating under. The result was a character with clear casual appeal: expressive, readable at mobile scale, and emotionally distinct from anything the early build had produced.
The boosters presented the same structural challenge in a different form — they existed only as rough product concepts with no visual execution. Rather than surface the gap as a blocker, I directed the team to design and animate them ourselves. The standard we held ourselves to wasn't placeholder quality — it was creative work that could plausibly inform the actual product direction. That distinction matters. Design decisions made at this stage have to be defensible on their own terms, not simply functional enough to fill a frame.
Getting the motion language right required real exploration across multiple dimensions simultaneously — and the decisions in each one affected all the others. The card flip dynamics were the mechanical heart of it. Cards are the primary interaction surface: how they move, resist, and reveal carries the core emotional weight of every play moment. We worked across a spectrum of energy levels — heavier dynamics with more physical weight on one end, faster and snappier movement on the other. Higher energy carried excitement but started to feel frantic, like the game was happening to you rather than being played by you. Lower energy felt considered but risked reading as slow. We were looking for something in between: tangible and satisfying, but light enough to feel effortless.
Character Direction
Rebuilt via AI-Assisted Concepting
Nano + Kling AI. Prioritised expressiveness, casual appeal, and readability at mobile scale — qualities the original 3D asset lacked.
Booster Design
Conceived and Animated In-House
Designed before product had defined them — held to a standard that could plausibly inform the actual product direction.
Gameplay Staging
Communicating Without a Working Build
Used motion hierarchy, cause → effect → payoff timing, and UI clarity to make the draw interaction feel intuitive on first watch.
Visual Language
Defined Live Across Production
Reward emphasis, composition logic, colour hierarchy, and pacing applied consistently across both video and store assets.