How AI Accelerated Game Asset Production by a Factor of 4

The REB8T Case Study: From Expensive Manual Production to a Streamlined Generation and Animation Process

client

REB8T
digital studio by Setters

https://rebooot.me/

Industry

Gamification, digital products, creative technologies, branding

Team size

<20 specialists

Example of use

Game mechanics, 3D assets, animations

Results

Game pieces

800+

Total number of game assets; currently in production: 832 icons and 60 animations

Speeding Up Production

4 times

150 hours using an AI pipeline instead of 600+ hours of manual work to create game assets

Challenge

The REB8T STUDIO team launched the “Alchemy” game mechanic within a banking app. Launching the feature required a large volume of assets in a consistent style: over 800 elements, including 832 icons and 60 animations. Using a traditional production process, this project would have taken too long and been too expensive.

"Without Neuron, a project like this would simply be too expensive and take too long."

— Vyacheslav Dyachkov, Lead UX/UI Designer at REB8T STUDIO

Scope of the Project and Timeline

For the “Alchemy” project, which involved a banking app, the team needed to create 832 3D icons and 60 animations. Using a standard approach, a project of this scale would have required hundreds of hours and pushed the budget to a level that would have made the launch unfeasible.

The Problem with the Classical Approach

Manual rendering and constantly switching between different tools were slowing down production. We needed to maintain a consistent style, make quick edits to elements, and coordinate work among several people simultaneously.

Implementation of Neural Networks

We need to build a simple, reproducible, and easy-to-understand pipeline that delivers predictable results: references, generation, spot editing, animation, and scaling on command without any loss of quality.

Decision

One platform.
Endless possibilities.

The team put together a simple and scalable pipeline: reference image → GPT Image to generate content in the desired style → Nano Banana for fine-tuning and refinements → Kling to animate selected elements. To conceptualize game mechanics, they used text models and quickly visualized ideas using Nano Banana.

GPT Image — Basic generation of 3D icons in the desired style based on reference images
Nano Banana — Selective Editing: Remove defects, add details, and modify elements without having to regenerate the entire model
Kling — animations of individual game elements for the product's final "wow" moments
ChatGPT / Gemini Text — Collecting and Organizing Prompts for Generating and Conceptualizing Game Mechanics

GPT Image

Nano Banana

Kling

ChatGPT / Gemini Text

"We needed to develop a very simple pipeline so we could quickly generate a large number of elements and easily bring new people into the process."

— Vyacheslav Dyachkov, Lead UX/UI Designer at REB8T STUDIO

Result

An AI-driven approach helped the REB8T STUDIO team transform a complex production process into a manageable one: shortening timelines, scaling up asset generation, and standardizing the style across a large volume of game assets.

1. The “Alchemy” Project: To launch the mechanics within the banking app, we needed to create 832 3D icons and 60 animations—a volume that could not be quickly produced using traditional methods.

Workflow: In 90% of cases, reference images were used, and the desired animation result was often achieved in 2–3 iterations using simple prompts in Russian.

2. Process efficiency: According to the team’s estimate, manual production would have taken 600+ hours; with the AI pipeline, the entire volume was completed in approximately 150 hours.

Concept Development for Sales: Ideas for game mechanics were visualized quickly—several concepts for a commercial proposal can be prepared in a single day.

3. Scaling to a team: The node-based structure in Phygital+ has simplified the onboarding of new members and enabled parallel work on a large pool of assets.

Watch Vyacheslav Dyachkov's presentation on the use of neural networks in gamification

"With Phygital+, the entire process is centralized in one place and clearly organized into nodes, so production moves significantly faster, and it's much easier to bring new people on board."

— Vyacheslav Dyachkov, Lead UX/UI Designer at REB8T STUDIO

Would you like to set up an AI-driven process for your product team?

We can help you build easy-to-understand AI pipelines for visual production, gamification, and design tasks—from pilot projects to scaling up to a team.