AI Brings New Coordination Layer to Streaming Ad Tech

AI is introducing a new coordination layer across the entire streaming ad monetization workflow. Media planning is shifting from static spreadsheets to living systems that can ask questions like "What if supply shifts mid-flight?" or "What if this audience over-indexes on live?" This enables planners to model delivery across live, VOD (video-on-demand), and FAST (free ad-supported TV) channels together and update forecasts as signals arrive, rather than planning channel by channel.

According to Luciano Marcos Escudero, VP of media engineering at Globant, 35–40% of publishers on the publisher side are currently using AI in ad tech. However, adoption is uneven. Jeff Ellin, VP of product architecture at FreeWheel, reports meeting with 18 European broadcasters and finding that some are eager to begin immediately, while others are still evaluating. Ellin notes that "buyers are doing their homework earlier, and they already have an LLM of choice that they're ready to use so they can connect to it and put their data through it a lot faster." Some publishers remain skeptical or uninformed; Ellin says 2025 is "more of a learning world," though eventually "AI is going to make the buying and the selling happen."

The industry is moving from basic machine learning into agentic workflows. The IAB Tech Lab recently unified AI standards under AAMP (Agentic Advertising Management Protocols), integrating modern AI protocols like Model Context Protocol (MCP) and gRPC into existing standards such as OpenRTB, AdCOM, and OpenDirect. Shailley Singh, COO and EVP of product at IAB Tech Lab, explains that "AI for the media ecosystem is creating this ability for advertising companies from brands to agencies to ad tech companies and publishers to create this super efficiency in how they value impressions, how they negotiate, and how they work and manage campaigns." IAB Tech Lab identifies three primary integration areas: autonomous transacting and execution (AI agents negotiating and executing deals dynamically), outcomes and measurement optimization (connecting upper-funnel engagement with lower-funnel conversions without relying on legacy-tracking pixels), and content and rights management (LLMs and AI systems negotiating in real time with publishers to crawl, index, and monetize content).

Research from IBM Institute for Business Value estimates that spending for agentic AI will triple in the next year, and 46% of those interviewed in Microsoft's 2025 Work Trend Index Annual Report said they are already using agents to fully automate workflows or processes. Two years ago, media companies were "a little bit resistant to introducing AI as part of the workflows," according to Globant's Escudero, but "now that is a much easier conversation." However, he notes that "there is not a workflow that we have identified so far that's completely automated. Most of the workflows that we're using with our customers are fully speeding up the process, but they're still having humans in the loop."

One significant use case involves AI understanding content metadata. Earlier workflows required planners to manually identify and differentiate assets like episodes and movies, then reach out to agencies—a process that could take weeks. Now, all new and old content comes with metadata extraction as part of the media ingestion workflow, including scene-level understanding of what types of artifacts and activities are present. Bitmovin launched AI Scene Analysis last year, which provides contextual metadata enrichment using multimodal AI models to extract and map content to IAB taxonomies. Jacob Arends, senior product manager for Bitmovin's AI Scene Analysis and Playback, gives an example: "If it's a car chase scene, your content taxonomies might be car, road, and traffic, but your ad opportunities might be automotive and travel, which can then be passed to the ad server, and that can then influence the decision of which ad is then served to the customer." Bitmovin charges 9 cents per input minute on a pay-as-you-go basis, and processing a single asset takes between 60 to 90 minutes, putting the total cost at "not more than $30–$35" per asset. However, Escudero notes that costs accumulate "when you multiply them by hundreds of assets," and the expensive part is processing an entire archive.

Another emerging use case is yield optimization and campaign management. By 2026, every publisher, SSP (sell-side platform), and DSP (demand-side platform) is expected to have a chatbot powered by various agents. FreeWheel's Ellin explains that clients are using agents to "find and identify the campaigns that aren't delivering yet" and to "tell me which programmatic deals are bidding or not bidding and help find and tweak those things a lot faster." What FreeWheel has found is that "clients are more comfortable getting those results and delivering them manually—that is, not requiring the agent to go off and actually make the change for them and trusting the results." Kantar worked with Microsoft to develop agents that turn decades of historical data into real-time, conversational insights, allowing media planners to ask detailed questions about reach, creative impact, and how to modify buys for better results. Globant's Escudero describes a hypothetical workflow: a planner interacts with a chatbot that understands what happened in movies or upcoming inventory and asks "give me different examples on the new inventory that is coming to our platform that will be able to make ad markers [for selling cars] across the movies." The chatbot responds with a list of movies and exact times where cars are discussed. Escudero notes that "most of our customers are not only using the context of the assets; they are also adding historical data from sales," applying both specific historical data and similar content types to a knowledge base of past sales. Dentsu built a forecasting and optimization agent to let media planners test assumptions and try out ideas.

The article describes a fundamental shift in how the ad tech industry operates. Until recently, media planning relied on static spreadsheets and manual processes that could take weeks—publishers had to identify specific content assets, reach out to agencies, and coordinate plans across different channels. AI is collapsing this timeline by enabling planners to ask dynamic questions like "What if supply shifts mid-flight?" and model delivery across live, VOD, and FAST channels simultaneously. The speed advantage is material: Globant's work shows that AI can cut asset identification from weeks down to the time it takes to understand metadata, which now happens automatically as content is ingested.

The industry's adoption curve varies significantly. While 35–40% of publishers are using some form of AI, Jeff Ellin from FreeWheel reports conversations with 18 European broadcasters, some expressing immediate readiness ("Can we start tomorrow?") and others still evaluating. Standardization—specifically through IAB Tech Lab's new AAMP framework—is emerging as critical infrastructure, extending existing protocols like OpenRTB rather than replacing them wholesale. This allows publishers and buyers to connect their own AI systems through Model Context Protocol (MCP), reducing friction and enabling natural-language interaction with ad platforms.

Cost and ROI remain open questions. Bitmovin's per-asset pricing ($30–$35) appears reasonable until multiplied across hundreds of archived assets, which is why publishers are conducting return-on-investment analyses. Notably, clients are not yet trusting agents to execute changes autonomously; instead, they prefer agents to identify underperforming campaigns and deliver recommendations for manual review. This suggests the technology is advancing faster than organizational readiness to delegate decision-making to machines.

No comments yet. Be the first to share your thoughts!