What challenges did Globo face before AI implementation?

Operators previously had to open each video, find the clapperboard frame, and manually transcribe program, scene, director, and timing details into the media system. The workflow was repetitive, labor-intensive, and slowed production and post-production operations.

What was the impact of the AI solution?

The workflow became fully automated. Google Cloud reported an 80% reduction in total processing time for the workflow. The solution covers more than 85% of the materials processed. Metadata filling accuracy was reported at nearly 100%. The company says operators are freed from repetitive manual work and can focus on higher-value tasks.

Rank#2,943of 3,449

Source

GCPExpandedEvidence: Medium50/100

Globo: Automated digital clapper metadata using Cloud Vision AI (OCR)

Use case typeMedical document automationUpdated Jun 13, 2026

Globo, the largest media company in Latin America, is using Google Cloud to modernize media production and post-production workflows. For clapperboard metadata, the company replaced manual transcription with an automated process that extracts clapper information from images and fills production templates. A Python application detects the clapper, creates a low-resolution proxy, sends the frame to Cloud Vision AI by API, and writes the returned metadata into an XML template from Cloud Storage before passing it to the media system.

Organization: Globo
Industry: Other
Location: Brazil
Published: January 2021

Reported outcomes

−80%

timeTime & speed

+85%quantified impact100%accuracy

Strategic outcomes

Speed & agilityFully automated clapper metadata workflowEmployee experienceFreed operators from manual workNew product / capabilityAutomated metadata extraction from clapperboards

Catalog median for time & speed deployments: −50% across 312 reported metrics. Compare benchmarks →

Why do we believe this?Outcome claims, sources, and evidence checks

Normalized claim

Time: 80% decrease

Google Cloud Customer StoryJan 1, 2021Customer storyInferred claimMedium evidence strength

Google Cloud reported an 80% reduction in total processing time for the workflow.

Normalized claim

Quantified impact: 85% increase

Google Cloud Customer StoryJan 1, 2021Customer storyInferred claimMedium evidence strength

The solution covers more than 85% of the materials processed.

Normalized claim

Accuracy: 100%

Google Cloud Customer StoryJan 1, 2021Customer storyInferred claimMedium evidence strength

Metadata filling accuracy was reported at nearly 100%.

Why do we believe this deployment?Customer identity, provider attribution, maturity, and source checks

Customer: Globo
Provider: GCP
Maturity: Unknown
Linked source: Google Cloud Customer Story

No explicit deployment-stage evidence found.

Customer identity supportedSource describes one deploymentMaturity evidence evaluated

Primary read

Use case focus

Showing 3 of 3

1Workflow Automation
2Document AI
3OCR

Operators previously had to open each video, find the clapperboard frame, and manually transcribe program, scene, director, and timing details into the media system.
The workflow was repetitive, labor-intensive, and slowed production and post-production operations.

The application generates a short proxy video containing the clapperboard frame from the original camera footage.
The proxy is uploaded to Cloud Storage and triggers a process that calls Cloud Vision AI OCR through an API.
Cloud Functions formats the OCR output, and the metadata is inserted into an XML template stored in Cloud Storage.
A web application for standardized clapper creation runs on App Engine with Firebase-based login configuration.
The process is implemented with a Python application and uses private API access for the AI call.

Technologies

Cloud Vision AI Cloud Storage Cloud Functions App Engine Firebase Compute Engine Cloud Interconnect

The workflow became fully automated.
Google Cloud reported an 80% reduction in total processing time for the workflow.
The solution covers more than 85% of the materials processed.
Metadata filling accuracy was reported at nearly 100%.
The company says operators are freed from repetitive manual work and can focus on higher-value tasks.

Architecture

A Python application extracts a clapperboard frame from camera footage, creates a low-resolution proxy, uploads it to Cloud Storage, and triggers Cloud Vision AI OCR via API. Cloud Functions formats the output, XML templates are retrieved from Cloud Storage, and a web application on App Engine with Firebase authentication supports standardized clapper creation. The solution uses private API access and does not send full video content to the cloud.

Sources & evidence1

Evidence: Medium50/100Evidence strength