# StreamTV Panel Identifies AI, Metadata, Personalization as Solutions > Source: > Published: 2026-06-18 20:02:22.257775+00:00 # StreamTV Panel Identifies AI, Metadata, Personalization as Solutions Media Play News reports that a June 18 StreamTV Show panel in Denver brought together executives from Tubi, Cineverse, Gracenote, Mediagenix and DirecTV to discuss content discovery. Moderator Tyler Aquilina of Luminate cited Gracenote research (November 2025) showing consumers spend an average of 14 minutes searching for something to watch. Cineverse President, Technology & Chief Product Officer Tony Huidor said "search is broken across the industry" and advocated moving from keyword-based to natural-language search. Tubi VP of Product, Discovery & AI Sharon Kritzer noted the company manages a library of 300,000+ titles (per Tubi) and said metadata plus AI-driven, context-aware personalization are central to improving discoverability. The panel emphasized richer metadata, natural-language search, and hyper-personalization as complementary approaches to reduce search friction, per Media Play News. ### What happened Media Play News reports that a June 18 StreamTV Show panel in Denver convened executives from **Tubi**, **Cineverse**, **Gracenote**, **Mediagenix** and **DirecTV** to discuss content discovery challenges and emerging technical responses. Moderator Tyler Aquilina of Luminate cited **Gracenote** research (November 2025) showing consumers spend an average of **14 minutes** searching for something to watch globally. **Cineverse** President, Technology & Chief Product Officer Tony Huidor said, "Part of the problem is search is broken across the industry," and argued the sector should move beyond keyword search to natural-language approaches. **Tubi** VP of Product, Discovery & AI Sharon Kritzer said the service manages **300,000+ titles** (per Tubi) and framed metadata plus AI as necessary to guide viewers through large libraries. ### Technical context Industry-pattern observations: panel remarks map onto two converging engineering efforts seen across streaming platforms. First, richer, structured metadata (genre facets, granular attributes, semantic tags) is becoming a prerequisite for effective retrieval and recommendation. Second, natural-language search and contextualized ranking layers leverage embeddings and retrieval-augmented methods to match queries to content signals beyond simple keywords. These are general trends; the panel comments reflect public-facing priorities rather than disclosed engineering roadmaps. ### Context and significance Reducing observed average search time requires both backend data investments and runtime personalization. Panels like this indicate that practitioners are prioritizing metadata enrichment pipelines, intent-aware models, and session-based personalization to improve conversion from discovery to viewing. That framing aligns with broader adoption of vector search, embeddings for semantic retrieval, and session-aware ranking in recommender stacks. ### What to watch For practitioners: watch for vendor announcements and engineering case studies detailing metadata schemas, annotation tooling, and the integration of embedding-based retrieval into existing recommender flows. Also monitor measurements beyond click-through - session duration, abandonment, and successful content matches - to evaluate discovery impact. ## Scoring Rationale Useful conference-panel report for streaming ML and recommendation-system practitioners, covering natural-language search, embedding-based retrieval, and metadata enrichment as active industry priorities. The 14-minute search-time figure (Gracenote, 2025) provides concrete grounding, but the story is panel commentary rather than a product launch or research finding, placing it solidly in the 'solid/niche' tier rather than 'notable'. Practice with real Ad Tech data 90 SQL & Python problems · 15 industry datasets [Active Search Campaigns by BudgetEasy](/problems/sql/active-search-campaigns-by-budget) [High CPC Clicks & Poor Landing PagesMedium](/problems/sql/high-cpc-clicks-poor-landing-page) [Campaign ROAS by Attribution ModelHard](/problems/sql/campaign-roas-by-attribution-model) 250 free problems · No credit card [See all Ad Tech problems](/problems/datasets/adtech)