Rippin' into AI
Rippin' into AI
October 21, 2026

Why I built this

Everyone’s talking about AI, and for creative professionals it’s a complicated conversation. I wanted to build something that shows I can bring AI into my workflow, and that my years of product experience let me use it to extend my skills, not replace them.

Reward moments are critical in games and in any product that uses gamification to engage users. A digital card pack opening felt like the right project to show what I can do.

The starting point

I started with a rough storyboard: a sealed foil pack, a swipe that tears it open, edges that burn into embers, building magic, and a final card reveal. I fed it into Cursor to see what came out.

The first version hit the beats of the storyboard, but the visuals were rough and generic. That’s where the real work started: taking a functional first pass and turning it into something that looks and feels like a finished product.

I did a rough storyboard of how I wanted the pack to open. Really rough.

Process: directing AI like a technical artist

I worked in Cursor over about 120 prompts, treating it like a fast technical artist. I judged every change in the browser and gave feedback the way I would in a VFX review: specific percentages, timing in seconds, and annotated screenshots marking what was wrong.

1. Making the burn believable

Most of my prompting went into how foil actually burns. I tuned it in layers, one at a time:

  • Singe: a long yellowing that runs ahead of the tear, darkening to brown and then a solid black rim behind the embers.
  • Burn line: a wide orange glow with a tight yellow-white core, plus bloom on the hottest areas.
  • Embers: flakes that start glowing orange, cool to charcoal and grey ash, and shrink fast early, then slowly as they die.
  • Tear shape: a jagged, straight-sided V that tears top to bottom and opens widest at the end, with the foil burning away in place instead of sliding off.

2. Building anticipation

I replaced the swipe with a slide-to-rip control, which gives the user a clear moment of commitment. Then I loaded the slider with anticipation:

  • The first stretch of the slider scrubs the start of the rip, so heat builds and the top starts to split under your thumb.
  • The pack shakes harder the further you slide.
  • Lightning crackles around the edge of the pack in short traveling bursts, faint at rest and faster as the slider moves.
  • Three layered energy rings orbit the pack at different speeds, then blow outward and fade when the rip commits.

3. Making rarity matter

I added Rare, Epic and Legendary tiers and set a firm rule: nothing hints at the tier until the card flips. Every pack looks the same until the reveal, which keeps the experience honest for any product built on fair odds.

At the flip, a tier-colored beam snakes around the card, the background tints, smoke blooms behind it, and the glow and particle count rise with rarity. I timed each tier’s flip from a rough motion guide I made in After Effects. Rare floats down light and feathery, Legendary lands with real weight, and Epic sits between them.

‍

The UI design system. I was pleasantly surprised by how well Cursor took my design system and populated it within the demo.

Art and design

The code handles motion and timing. The look comes from assets I made by hand, each one dropped into the project folder and wired in by name.

  • Pack, card back and card front art (Photoshop): the swappable art, with alpha for the cards’ rounded corners.
  • Smoke particle sprite (Photoshop): wispy smoke as the pack opens and behind the revealed card.
  • Lightning crackle frames (Photoshop): the traveling crackle around the sealed pack.
  • Energy ring textures, 3 layers (Photoshop): orbiting rings that build as you slide.
  • Rarity beam textures, 3 tiers (Photoshop): the beam that wraps the card on reveal.
  • Slider fill sprites, 3 layers (Photoshop): scrolling fill textures that speed up as you slide.
  • Icons (Illustrator): the Sealed, Rip, Burn, Spin and Pull timeline.
  • UI Kit and Layout pages (Figma): components, colors, type and placement for the whole interface.
  • Concept frames and paint-overs (Photoshop): visual direction for the burn, the tear shape, the beam and the rings.

The concept frames did the most work. When a description wasn’t landing, a quick paint-over (green lines showing the tear angle, a sequenced crackle concept, a marked-up singe) got the next iteration right far more often than more words did.

Figma became the source of truth for the interface. I built a UI Kit page for components and a Layout page for placement, and the prototype matched them to within a few pixels.

‍

I created the VFX sprites from scratch in Photoshop.

Outcome and Impressions

The result is a live, mobile-ready pack opening with a burn I’d put next to a shipped game, rarity-driven reveals, and art that swaps by file. It went from storyboard to a live site in six days.

The bigger result was the workflow. I directed the build the way I’d direct a VFX artist and a front-end engineer, made the assets myself, and shipped without waiting on anyone.

More Thoughts