How Ortana & DeepVA are transforming media and broadcast companies with superior AI
In order to streamline these processes, organisations require reliable automation, consistent metadata and seamless integration between scheduling, storage, review tools and AI analysis.
The Core Challenge
Media companies often face recurring questions when planning AI innovation:
- How can repetitive post-production tasks be automated without losing creative control?
- How should post-production from scheduling to cloud review work together across several tools?
- Can we have all necessary steps consistently for every episode or project?
- How do we ensure predictable, repeatable workflow execution?
Automation is only effective when all components work together seamless.
Solution Overview: Cubix as Orchestration, DeepVA as the AI Engine
With Ortana’s Cubix media-aware workflow engine (MAWE) with its 160+ native API integrations and DeepVA’s AI modules, organisations can build automated post-production pipelines for any repetitive or episodic content. Cubix providing the glue between all solutions involved:
Cubix orchestrates:
- DeepVA for visual, audio, and contextual analysis
- Freispace for project creation, planing and suite scheduling
- Iconik or Frame.io for collaboration and review
- storage (cloud/on-premise), ingest, edit platforms, and many more possibilities
DeepVA analysis (via Cubix integration):
- automatic shotlists and scene boundaries
- face recognition with custom datasets of your cast
- object, concept and context detection
- speech-to-text, speaker detection, subtitle drafts
- generating finished draft timeline by utilizing a safe & local-hosted LLM
Example Workflow
Project Creation (Freispace)
A new project triggers automatic suite booking and folder creation.
Cubix sets up the workspace online as well as offline.
Ingest & Proxy Creation
Incoming media is detected by Cubix, which automatically generates proxies and prepares assets for editing and reviewing.
AI Analysis (DeepVA)
Cubix sends media to DeepVA for:
- shotlist automation
- character/face identification
- transcript generation
- visual metadata extraction
All results return in a standardised, timecoded format, guaranteeing a predictable and consistent output for storage as well as post-production.
Automated Rough Assembly
Using this metadata, DeepVA can generate a draft edit / first assembly or bin, by combining the knowledge about your needs with the knowledge and metadata of your assets. All being processed on a local and secure LLM, without cloud based analysis.
Templates ensure consistency across similar projects or episodes.
Review & Collaboration (Iconik or Frame.io)
Cubix publishes the first assembly or proxies to Iconik or Frame.io for immediate review. Comments and approvals can trigger further automation.
Compliance & Reliability
DeepVA and Cubix provide a tightly integrated compliant, scalable, and future-ready AI foundation for media companies. Regulatory requirements, data security and AI Act principles are built directly into the API, reducing legal and operational risk. Continuous model improvements and predictable orchestration eliminate internal IT overhead, while a broadcaster-proven user experience ensures easy adoption in editorial and archive environments. With flexible integration of new models and features, broadcasters remain fully prepared for future technological and regulatory developments.
Strategic Impact
This integrated approach enables faster editorial decisions, consistent metadata, scalable automation and the transformation of regional knowledge into operational AI models — maintaining regional identity while improving production efficiency.
Interested?
Let’s talk!
