AI search overviews have reduced scope for discovery. But, says Shannon Millard of Epsilon, immersive channels such as CTV and audio still offer plenty of opportunities.
With AI search referrals driving less than 1% of traffic this year, streaming and audio offer hope, explains Millard (Rakhmat Suwandi/Unsplash)
AI-driven search is transforming how people find information. Already, 72% of people say they use AI tools for search. Immediate answers are a gratifying and novel response to the somewhat dated list of web links, but AI overviews strip away much of the exploration and discovery that defined traditional search. For marketers, that shift narrows the role search can play in reaching new audiences.
By contrast, immersive content channels such as connected TV and audio remain largely unaffected. People consume them for escapism and entertainment rather than information. These are environments where attention is plentiful, yet at the same time, advertising often fails to connect. That creates an opening to move beyond the intense competition around keywords, provided marketers are willing to reorient their approach from context-led to people-led.
By the numbers, AI-driven search referrals account for less than 1% of traffic this year. The statistic positions AI search more as a research channel, while organic search drives conversions and growth. If research-led search is fading, discovery must come from elsewhere.
Advertisement
Streaming and audio are natural destinations. Viewers and listeners are already in an engaged state of mind with every film, show, or podcast posing moments of intent – holiday ideas, food inspiration, fashion cues – where advertising could be both relevant and welcome.
But all too often, the ads they receive are repetitive and generic, because spend remains tied up in legacy publisher deals that lack the advanced targeting many expect. Reaching real people in those moments with thoughtful creative is an opportunity that is still going wanting.
Traditional media buys primarily follow contextual signals. Someone searching for London Marathon articles, or watching highlights of last year’s event, is among the most likely to be in the market for running shoes. But they could actually be researching travel tips to cheer on a family member, or even checking the route to make sure their weekend plans won’t be disrupted. Context alone cannot capture the true intent. Immersive formats require a deeper understanding of the individual.
For advertising in immersive channels to work, relevance must center on the person. Campaigns should be oriented around audiences genuinely in-market, using transactional and loyalty data to connect behaviors with interests. This is why CTV and retailer partnerships are so powerful today; dynamic personas from real-time purchases, loyalty data, and engagement data keep campaigns fresh and relevant. That means a food brand can target shoppers who recently bought pasta directly within a cooking show, then measure the uplift at checkout.
The most immediate opportunity actually lies in predictive rather than generative AI. When paired with person-first identity resolution, it allows marketers to deliver contextually rich messages that respect the immersive experience. Imagine a podcast ad that mimics the sound of noise-canceling headphones for a listener who has already placed earbuds in their basket but not yet checked out.
Adding context to this equation can make the placement even more powerful, like running the ad during rush hour, when the listener is on a crowded train. It’s hyper-relevant, illustrates the product benefit in real time, and complements the moment rather than interrupting it.
The other benefit of person-first identity is measurement. When campaigns are anchored in verified profiles rather than proxies, it becomes possible to close the loop between exposure and outcome. That means knowing whether a CTV ad drove a sale in-store, or whether a podcast placement shifted long-term loyalty. Measurement here is not about vanity metrics but real business growth.
Advertisement
This people-first approach is also the foundation for the next wave of AI tools. AI agents, designed to manage campaigns and interact with consumers, will grow in use as capabilities and interoperability expand. Appetite is already clear: only a small minority of senior marketers say their organizations will never permit AI to make independent decisions, while nearly half of those developing in-house models have already granted some level of decision-making authority.
This trajectory underlines why strong, flexible data foundations matter today. Without refined and scalable first-party data, AI agents and predictive tools alike will deliver shallow results. With them, marketers can enable seamless integration with AI-driven platforms, unlock new levels of efficiency in media buying, and scale campaigns that remain rooted in a genuine understanding of individuals.
So, while some focus on improving hit rates for context-specific AI advertising, the brands that will succeed are those investing in understanding individuals. Optimization starts with people, and AI follows.
Epsilon
Epsilon is a global advertising and marketing technology company positioned at the center of Publicis Groupe. We connect advertisers with consumers to drive performance...
