Seedtag’s Brian Danzis on scaling contextual advertising in a fragmented CTV ecosystem

As connected TV continues its rise as the dominant viewing platform, advertisers are under growing pressure to deliver relevant, privacy-compliant messaging without relying on traditional identifiers like cookies or device IDs.

Contextual AI is a viable solution to the intricate problem of targeting – an approach that uses artificial intelligence to analyze the content of a page or program (text, images, video and metadata) to determine its meaning, tone and themes, allowing ads to be placed in environments that align with a brand’s messaging without tracking the viewer.

Seedtag, a contextual advertising company now active in 17 markets, is expanding its approach into CTV. Known for its privacy-first targeting capabilities on the open web, the company recently acquired Beachfront (a sell-side ad platform) to build direct publisher integrations and establish its foothold in the streaming space.

“We’ve experienced a renaissance in contextual targeting, initially because of signal loss and privacy concerns,” said Brian Danzis, president of North America at Seedtag, in an interview with NCS. “But now it’s also about enabling advertisers to reach more nuanced audiences in ways other strategies can’t.”

Connecting signals across screens

The shift toward contextual targeting is part of a broader reckoning in adtech. With cookie-based targeting on the decline and streaming platforms limiting access to granular user data, brands are turning to strategies that emphasize content alignment over identity.

“CTV has matured rapidly, but it’s still not a unified ecosystem,” said Danzis. “The absence of consistent identifiers forces advertisers to think differently about relevance, and contextual is proving to be a scalable way forward.”

CTV fragmentation has created a scale paradox. While adoption is near-universal in the U.S., with Danzis citing more than 600 video apps available on a single Samsung TV model, the sheer volume of content makes targeting more complicated. Most CTV platforms offer only genre-level segmentation, missing the more specific contextual signals marketers increasingly demand.

Seedtag’s platform, powered by its proprietary AI engine called Liz, analyzes content using natural language processing, computer vision and sentiment analysis. It breaks down articles and imagery into over 11,000 categories – beyond the standard IAB taxonomy. Visual data is equally detailed, with Liz capable of identifying 600+ objects and 300 distinct situational contexts.

Advertisement

Now, the company is applying these capabilities to video, including CTV content. By analyzing both what users read and what they watch, Seedtag aims to unify signals across the open web and streaming environments.

“If someone reads articles about sustainably sourced clothing, and then watches streaming content tied to ethical fashion, our platform can surface that connection,” Danzis explained.

While scene-level contextual targeting in CTV is still limited, Danzis said the company is building the infrastructure to get there.

“This isn’t years away – it’s quarters,” he noted.

Moving beyond keywords and age brackets

Danzis acknowledged that contextual targeting still carries misconceptions, often being equated with simple keyword matching. But keyword systems can miss nuance or produce false positives. Seedtag addresses this by training its AI across multiple languages and dialects, enabling it to understand colloquial phrases, slang and region-specific context.

The Liz platform allows advertisers to upload a detailed campaign brief, which the system then uses to identify relevant articles, imagery and keywords across Seedtag’s network. An “affinity index” lets brands fine-tune how closely aligned the content should be to their message. Results can be used not only for targeting but also for optimizing creative, search strategy and channel selection.

“We don’t just show advertisers where their content is running – we also surface opportunities they might be missing,” Danzis said. That includes identifying underserved sports segments, emerging sustainability topics or culturally relevant trends that align with a brand’s values.

Publishers also benefit from this model. By breaking broad categories like news, sports or lifestyle into specific themes, Seedtag helps publishers better monetize long-tail content through precise packaging and targeting.

Targeting without identifiers

Seedtag’s contextual approach is designed to function independently of user-level identifiers. This privacy-first stance is becoming increasingly attractive as regulations tighten and large platforms limit third-party tracking.

“Marketers want precision, but they also want to stay compliant – and contextual enables both,” Danzis said.

On CTV, concerns about spoofing and viewability persist. To address this, Seedtag has partnered with firms such as Human and Jounce Media and uses proprietary technology to ensure ad delivery occurs only when a TV is actually on and playing approved content.

Measurement and attribution are also key areas of investment. Seedtag integrates with partners to track metrics beyond impressions – looking at brand lift, purchase intent and even physical foot traffic.

Advertisement

The vision is to make Liz a day-zero tool, capable of ingesting an agency’s brief and offering targeting strategies or cultural hooks before the campaign even launches.

Danzis believes CTV will soon move further into performance territory, driven by DTC brands and local advertisers accustomed to granular targeting.

“There’s an evolution underway,” he said. “We’re already seeing more pressure for attribution and outcome-driven buying.”

As media buyers become more sophisticated, they’re asking more thoughtful questions – and holding vendors accountable for performance, not just presence.

“We’ve moved beyond view-through rate as the only KPI,” he said. “Buyers want to know where their ads run, how often, and what the actual impact is.”

For Seedtag, contextual AI is not just about placement, it’s about better alignment with viewer values and preferences, especially as marketers look beyond traditional demographics.

“We’ve found that age and gender targeting often reinforces outdated assumptions,” Danzis said. “With contextual, we can reach audiences based on what matters to them – what they read, what they watch, and what they care about.”