Netflix: AI-Powered DOOH Rendering

Overview

Automate Digital built an AI-driven 3D pipeline for Netflix's digital out-of-home (DOOH) advertising that batches renders, resizes them to each screen specification and upscales them for large outdoor formats. The result was a 60% cut in render delivery time on out-of-home projects.

The Client

Netflix needed hundreds of out-of-home advertising renders for regions around the globe. The work shown on this page includes 3D anamorphic projections for Money Heist and Red Notice, filmed in situ.

The Challenge

Traditional rendering workflows were time-consuming, with each render taking days or weeks to complete. The manual process of resizing, reformatting, and upscaling for different screen specifications was creating bottlenecks in their advertising campaigns.

What is DOOH anamorphic 3D?

Digital out-of-home (DOOH) advertising runs on digital screens in public places. Anamorphic 3D content is made for one specific screen and viewing position. The image is distorted on purpose, so that from the intended viewpoint it appears to come out of the screen. Because each screen has its own size, shape, resolution and geometry, one creative idea can need many separate renders. When a campaign runs in many regions, the number of versions grows quickly, and so does the manual work.

Money Heist 3D Anamorphic Projection

Our Approach

We treated the repeated steps after each creative render as a pipeline rather than as separate manual tasks. Batching, resizing to screen specifications, reformatting and upscaling follow rules that can be written down. When those steps are defined once, a system can apply them to every render, for every screen, in every region.

The general aim of this kind of pipeline is to keep people on the creative work and move the repetitive delivery work to automation.

The Solution

We built an AI-driven 3D pipeline that changed their rendering workflow. The system automatically batches renders, intelligently resizes content to match specific screen specifications, and upscales content for massive outdoor formats without quality loss.

Key Features:

Intelligent Batch Processing: Automatically queues and processes multiple renders simultaneously
Smart Resizing: Adapts content to various screen specifications without manual intervention
AI-Powered Upscaling: Maintains visual quality when scaling to massive outdoor formats
Regional Optimization: Automatically adjusts content for different geographical markets

Red Notice 3D Anamorphic Projection

The Results

60%

Faster Delivery

Cut render delivery time on out-of-home projects by more than half

100+

Renders Per Week

Increased output capacity to handle global campaigns efficiently

90%

Manual Work Reduced

Eliminated repetitive manual tasks in the rendering pipeline

Technologies Used

3D Rendering AI Upscaling Batch Processing Automation Pipeline Cloud Computing Python

Lessons for Other Agencies

Separate creative work from delivery work. The creative idea is made once. Versions for each screen, size and region are delivery work, and delivery work is usually rule-based.
Write down the screen specifications. A pipeline can only resize and reformat correctly when each output format is defined clearly.
Automate in batches. When many renders follow the same steps, process them as a batch instead of one at a time.
Measure delivery time before you start. A baseline lets you show the change in turnaround after automation.

Frequently Asked Questions

What did Automate Digital build for Netflix?

An AI-driven 3D pipeline for out-of-home advertising renders. It batches renders, resizes content to each screen specification, upscales content for large outdoor formats and adjusts content for different regions.

What is anamorphic 3D advertising?

It is content made for one screen and one viewing position. The image is distorted so that, from that position, it appears to be three-dimensional. Each screen needs its own version.

What results did the pipeline deliver?

A 60% cut in render delivery time on out-of-home projects, output of 100+ renders per week and a 90% reduction in manual work in the rendering pipeline.