CASE STUDY
When Scale Meets Intelligence: How Protege and TwelveLabs Transformed Video Data Delivery
The Challenge
When foundation model companies come to Protege, they don't ask for simple things.
A request could be along the lines of: "We need thousands of video clips of high camera motion” or “thousands of clips of wind blowing through trees in nature.” . And we need it in two weeks, not two months."
This is where Protege's unique position as the single source of video content and AI data expertise becomes critical for delivering these novel data requests. Unlike single rights holders who can only provide a small subset of content, Protege aggregates content across the entire spectrum, giving customers the breadth and depth that meets both the broadest and narrowest AI training and evaluation demands.You need hands and feet? You don't want just soccer footage. You need sampling across all types of content such as sports, dance, medical content, everyday activities, and much, much more. That's what Protege delivers.
But even with access to vast, diverse archives, a fundamental problem remained: how do you find what you need across tens of thousands of hours, with precision and speed?
The TwelveLabs Advantage
This is where symbiosis matters.
Protege brings aggregated access to diverse, high-quality content from high-quality data sources that have been ethically licensed specifically for AI use cases under a single, straightforward licensing agreement, along with AI data-specific know-how and expertise from its researcher team. This provides premium content and AI knowledge that individual rights holders cannot provide at scale.
TwelveLabs brings multimodal video intelligence, AI models that actually understands what's happening in video, not just what tags say.
Together, they were able to jointly solve the problems of content scale and precise content creation. Importantly, this took days, rather than quarters
How It Works
Protege's Unique Data and AI Data Expertise:
Aggregates diverse content sources single providers can't match
Provides both scale (volume) and depth (diversity) simultaneously
Applies rigorous research and curation processes beyond raw search
Delivers compliant, rights-cleared, production-ready datasets
TwelveLabs' Marengo Model:
Natural language search across visual, audio, and motion patterns
Semantic understanding that captures nuance manual tagging misses
Intelligent segmentation producing precise 10-20 second clips
Structured JSON outputs ready for immediate use
The Combination:
Customer specifies complex, idiosyncratic video requirements
Protege identifies relevant content across aggregated sources
TwelveLabs' AI extracts precise moments matching specifications
Protege applies additional curation and data pipeline processes to ensure quality and compliance
Delivered in weeks, not the quarters or years that traditional approaches typically require
The Results
Two weeks. That's how long it took to deliver what would typically need 6+ months.
인덱싱 한도
True scale: Tens of thousands of hours processed with precision
Quality assurance: Combined AI intelligence + human curation
Seamless scaling: Infrastructure that grew as project scope expanded
Why This Partnership Works
Better Together
Together, Protege and TwelveLabs were able to create entirely new capabilities not previously available for both content rights holders and AI labs:
Protege's aggregation solves the content diversity problem
TwelveLabs' intelligence solves the precision and speed problem
Combined curation ensures quality that pure automation can't achieve
The market is evolving. Foundation model companies don't want generic TV libraries anymore, they want targeted, curated datasets delivered at speed. That requires both scale of access AND intelligent extraction — both are required for the world’s AI labs-
The Competitive Edge
Traditional approaches fail because:
Single rights holders lack diversity
Manual curation doesn't scale
Generic automation misses nuance
Timeline measured in quarters kills deals
Protege + TwelveLabs delivers:
Aggregated content breadth
Intelligent precision extraction
Professional curation quality
Speed measured in weeks
For Organizations with Massive Video Needs
If you're a foundation model company or AI developer with complex video data requirements:
You don’t have to settle for single-source content lacking diversity
You don’t have to accept quarter-long timelines in a fast-moving market
You don’t have to compromise between scale and quality
The combination of Protege's unmatched content aggregation & AI data expertise along with TwelveLabs' video intelligence creates something neither could build alone, and something the market increasingly demands.
The future belongs to organizations that can deliver both scale and intelligence, quickly. The tools exist today. The question is whether you'll move fast enough to leverage them.
