Twitter Versus AI Game

The Challenge

Display advertising had run into a wall of banner blindness. On feeds like Twitter, static ad units get scrolled past before they register, and the usual response, bigger, louder, more intrusive, only deepens the fatigue. The brief was to design something categorically different: Versus AI, a social ad format that trades passive impressions for active play. Instead of watching an ad, users tap into a quick, competitive, head-to-head game built from the partner's own content. The core design problem was making that game feel native to the feed, instantly understandable, and worth a tap in the half-second before a thumb keeps moving.

The harder constraint was scale without sameness. The format is powered by proprietary AI that transforms existing partner content into the Versus experience and drops it into each partner's content ecosystem, meaning the interface could not be a hand-crafted one-off. It had to be a system: a flexible brand and UI framework that stays recognizably 'Versus' while absorbing wildly different source material across partners like Twitter and Microsoft. Every screen had to hold up as a template an AI could populate, not a static comp that only works once.

Approach

1. Define the Versus mechanicBefore any pixels, I pinned down the core interaction, a pick-a-side, head-to-head matchup that reads in seconds and rewards a quick, competitive choice. Nailing the mechanic first kept every later screen in service of the game rather than decoration around it.
2. Make the entry point feed-nativeI designed the ad unit to belong in the feed instead of interrupting it, using low-friction tap-to-play framing so the leap from scrolling to playing costs the user almost nothing. The goal was an invitation, not an interception.
3. Map the full game flowI storyboarded the experience end to end across five screens, from the initial prompt, through the matchup and the user's choice, to the result and the moment worth sharing. Each screen carries one job and hands cleanly to the next.
4. Build a flexible brand systemBecause the AI generates instances from different partners' content, I built a brand and UI system with clear rules for type, color, and layout, flexible enough to flex per partner, disciplined enough to always read as Versus.
5. Package it for AI templating and handoffI delivered the whole experience in Figma as a structured component library with defined states and specs, so the proprietary AI and engineering partners could populate and ship instances without redrawing the format each time.

Key Features

Results

• Delivered a complete five-screen Versus game flow, from feed entry through the competitive matchup to the shareable result.
• Established a flexible brand and UI system that lets the proprietary AI generate on-brand Versus instances across different partners.
• Packaged the full experience as a state-aware Figma component library, ready for AI templating and engineering handoff.
• Shaped a reusable, feed-native ad format positioned to run inside multiple content ecosystems, including Twitter and Microsoft.