Comcast Technology Solutions is expanding its VideoAI framework with six AI-powered applications designed to automate video production and distribution tasks across broadcast, streaming and sports, including localisation, intelligent clipping, metadata generation and automated quality control.
The applications will integrate with Comcast's Cloud Media Processing orchestration layer, allowing media companies to introduce AI functions before, during or after existing processing workflows.
Commercial availability will be phased through 2026 and 2027, with Comcast set to demonstrate the expanded capabilities at IBC2026 in Amsterdam from September 11 to 14.
Six AI applications target different stages of video production
The expanded VideoAI portfolio covers video verticalisation, regional-language localisation, Smart Chapters, Metadata Enrichment, Automated QC and Intelligent Clipping.
Rather than operating as isolated AI tools, the applications are designed to sit within Comcast's existing video processing infrastructure.
Integration with Cloud Media Processing means broadcasters, streaming platforms and rights holders could introduce AI automation without having to create completely separate workflows for individual tasks.
The applications will also be available through Comcast Media360, its managed video supply-chain platform, Comcast Sports360 for sports rights holders and its standalone Cloud Video Platform.
AI automatically converts landscape video for vertical platforms
The video verticalisation application automatically reformats landscape content for portrait viewing.
The system detects where the main action is taking place and can introduce cuts and pans to keep relevant subjects within the vertical frame.
This addresses a growing production challenge for broadcasters and content owners distributing traditional widescreen footage to mobile-first services and social platforms.
Automating the process could reduce the amount of manual editing required to repurpose long-form or live content for vertical formats.
Localisation tool combines translation, dubbing and captions
Comcast's regional-language localisation application is designed to automate several stages of preparing content for different language markets.
The technology extracts dialogue, translates it and preserves background sound and music while generating new dubbed speech.
It also supports AI-generated voice matching and captions.
Combining these functions could simplify workflows that traditionally require separate transcription, translation, dubbing, audio processing and subtitling stages.
For streaming platforms and international broadcasters, faster localisation could also make it commercially viable to distribute a larger content catalogue across regional markets.
Smart Chapters identifies advertising opportunities
Smart Chapters uses AI to analyse programmes and identify structural elements such as introductions, closing credits and potential advertising breaks.
Possible advertising positions can then be scored before downstream conditioning, transcoding and packaging processes are triggered.
This gives the technology both a viewer-experience and monetisation role.
Automating chapter detection can help platforms improve navigation through long-form content, while identifying suitable advertising positions could support more efficient preparation of content for ad-supported streaming models.
Metadata Enrichment enables natural-language video search
The Metadata Enrichment application analyses individual scenes and complete assets before feeding the resulting information into Comcast's Cloud Video Platform.
That data can then support semantic and natural-language discovery.
Instead of requiring viewers or operators to search using exact titles or predetermined keywords, the system is intended to retrieve content according to the meaning and intent behind a query.
Richer scene-level metadata could also support recommendation systems, content management, advertising and archive discovery.
For media companies managing very large video libraries, automating metadata creation could make previously difficult-to-search footage significantly easier to discover and reuse.
Automated QC monitors video integrity continuously
Automated QC uses AI to identify technical problems affecting video content.
The application continuously monitors for failures including black screens and missing audio, alongside other content-integrity issues.
Quality control has traditionally required combinations of automated rule-based systems and manual review.
Applying AI to the process could allow media companies to detect a broader range of faults while reducing the volume of content requiring human inspection.
The value will depend on how accurately the system can distinguish genuine failures from intentional creative elements within programming.
Natural-language prompts generate video clips
Intelligent Clipping allows users to describe the moment they want in natural language.
The system can then identify the relevant footage and create a clip for applications including sports highlights, social media content or trailers.
The capability could be particularly useful in live sports, where the commercial value of a highlight can depend on how quickly it reaches digital and social channels after an event occurs.
It could also reduce the need for production teams to manually search lengthy programmes or sports feeds for individual moments.
Comcast pushes VideoAI towards production workflows
VideoAI already supports functions including contextual advertising, content segmentation and live-to-VOD preparation.
The six new applications broaden its role across the media supply chain.
Bart Spriester, senior vice-president and general manager of streaming, broadcast and advertising at Comcast Technology Solutions, said the video industry is moving beyond hypothetical discussions about AI towards applications capable of addressing production problems.
The phased rollout through 2026 and 2027 indicates that the full set of capabilities will not become commercially available simultaneously.
Source: MEA Tech Watch Press Reporter
