AI’s effect on linear TV and streaming

John Ebbert

AI's effect on TV

Think AI talk stopped with the IAB Newfronts? Think again. AI is top of mind at the TV Upfronts, too, which kicked off this week in New York City.

CNBC’s Lillian Rizzo reported (lightly edited):

“NBCUniversal Chairman, Global Ads and Partnerships, Mark Marshall noted that AI has been ‘leveling the playing field between linear and streaming.’

The tech is helping to gather viewership, engagement and other data faster for linear TV.

‘Now we can actually show that this path of linear and streaming together is better, and prove that against marketers’ objectives,’ he said.Several advertising heads told CNBC they expect AI to take center stage at upfronts this year.

‘AI is really helping people gain visibility and be very, very strategic and smart about what they commit to in the upfront,’ said EDO CEO Kevin Krim.”

Read: “At TV upfronts, AI is in and corporate shuffles are reshaping the lineup” (May 11) – CNBC

Related: “TV? What TV? The 2026 Upfronts Will Be Heavy on Creators, Football and AI Ad Tech” (May 11) – The Hollywood Reporter

LLMs & CHATBOTS

Developments

  • OpenAI launches the OpenAI Deployment Company to help businesses build around intelligence (May 11) – OpenAI
  • Interaction Models: A Scalable Approach to Human-AI Collaboration (May 11) – Thinking Machines
  • Introducing Claude Platform on AWS: Anthropic’s native platform, through your AWS account (May 11) – Amazon Web Services

BRANDS

Gartner: What CMOs say about AI

Marketing Dive covered the latest Gartner CMO survey.

Highlights included:

  • “70% of CMOs say becoming a leader in artificial intelligence is a key goal for 2026, but just 30% feel their organizations have the infrastructure necessary to do so.”
  • “15.3% of marketing budgets on average are directed toward AI, however, organizations that are better equipped to scale the technology allocate an average of 21.3%.”
  • “Marketing budgets have remained relatively flat, with 7.8% of company revenue going to marketing in 2026, compared to 7.7% in 2025. However, organizations with optimized AI programs tend to receive a higher share of revenue, averaging 8.9%.”

Read: “AI remains a top priority for CMOs, but spending lags: Gartner” (May 11) – Marketing Dive

More: “Gartner 2026 CMO Spend Survey Finds CMOs Allocate 15.3% of Marketing Budgets to AI, But Only 30% Are Ready to Scale AI Capabilities” (May 11) – Gartner

From tipsheet: Seems like every CMO wants “in” on AI. But, can their organization handle the transformation required? Enter: AI change management.

FINANCIALS

Earnings reported yesterday

Viant

  • Markets: Viant shares were up 12.3% in after-hours trading yesterday.
  • Stat: “‘We are off to a strong start to the year, delivering contribution ex-TAC growth of 18%, exceeding the mid-point of our guidance, and adjusted EBITDA growth of 81%, exceeding the high end of our guidance,’ stated Larry Madden, CFO of Viant.”
  • Results: Viant Technology Announces First Quarter 2026 Financial Results (May 11) – Viant
  • Earnings call transcript: Viant Technology’s Q1 2026 call transcript – Investing[dot]com
  • Earnings call quote: “Coming soon, ViantAI and our Outcomes product will lever TVision’s attention data to optimize bidding decisions toward inventory that commands high attentive scores and features attractive attention-adjusted CPMs, driving yet another enhancement to return on ad spend efficiency that no other platform can offer.”

LLMs & CHATBOTS

Time to citation: How long does it take?

Josh Blyskal, co-founder and product leader of answer engine optimization (AEO) firm Profound, shared new data related to a question on the tip of every marketer’s tongue these days.

He said on LinkedIn:

“‘How long does it take to get cited in ChatGPT or Claude?’

This is one of the biggest open questions in AEO. I dug through billions of logs and pulled ~900 newly published marketing pages to track time-to-citation, here’s what’s up:

  • Median time from publishing to citation: 6.81 days
  • 75th percentile (75% of pages are cited within): 18.68 days
  • 90th percentile (90% of pages are cited within): 37.10 days

As far as I can tell, this is the first number on time-to-citation for either engine, especially Claude.”

Read more on LinkedIn. (May 11)

AEO versus SEO

AEO operates differently than search engine optimization (SEO).

Though not exactly an apples-to-apples comparison, a study last year by SEO firm Ahrefs found when it comes to traditional search engine results pages (SERPs), “72.9% of pages in Google’s top 10 are more than 3 years old (up from 59% in 2017).” That wouldn’t be possible given how recently AI chatbots emerged.

On the other hand, Ahrefs also found that “40.82% of pages that ranked in the top 10 did so within 1 month,” which aligns more closely with Profound’s findings around “time to citation” velocity — though ranking in SERPs and being cited in AI-generated answers are fundamentally different behaviors.

Read more from Ahrefs. (May 2025)

From tipsheet: The optimization dynamics inside AI chatbots may evolve differently from traditional SEO. Conversational systems — and their developers — have strong incentives to prioritize perceived usefulness and conversational quality because a bad conversation can quickly drive away the user. In traditional search, a bad result after a short keyword query is frustrating but rarely fatal to the experience.

That said, AEO will almost certainly develop its own version of “black hat” optimization tactics over time.

LLMs & CHATBOTS

Watching ad load for ChatGPT ads

CEO Alex Halliday of AI search optimization company AirOps said OpenAI’s ChatGPT is still in a testing phase when it comes to ads.

He provided an anecdote yesterday from members of his team:

“While our Content Engineers started using [ChatGPT ads], we went and looked at what is actually being served.

tl;dr we are early:

  • April 7 through 14: roughly 2,813 ads per 100,000 ChatGPT searches. Around 3% of queries returned a sponsored card.
  • April 15 onward: 47 per 100,000. The drop happened overnight. It has held flat for three weeks since.

The drop is noteworthy but not unusual given the transition from testing to self-serve at scale. My hypothesis is they were testing behavior at scale before shifting to production monetization.”

Read more on LinkedIn. (May 11)

From tipsheet: The measured ad volume is for the free and $8/month “Go” tiers of ChatGPT. Over time, one major question for OpenAI will be how aggressively it increases conversational ad load while preserving the utility and user experience that made ChatGPT valuable in the first place. If OpenAI goes public anytime soon, quarterly monetization pressure becomes another variable in the equation.

CONNECTED TV

Amazon taps AI for Prime Video optimization

Amazon rolled out new AI-powered creative optimization capabilities called Dynamic TV Creative. The company describes it as “our first capability to automatically personalize Interactive Video Ads (IVA) on Prime Video series and films based on viewer shopping behavior.”

Read more from Amazon. (May 11)

AdExchanger’s Victoria McNally interviewed Fabrice Rousseau, director of advertising experiences at Amazon:

“‘Advertisers want to have control over their own creative and feel as if they’re “in the loop,”’ Rousseau explained. At the same time, they’re willing to give Amazon autonomy to decide where, when and to whom that creative should be shown.

The goal is that viewers will end up seeing the right ad at the right time, therefore driving engagement with interactive elements and, ultimately, more sales.

Rousseau called the tool a ‘no-brainer’ even for AI-hesitant brands, because it’s rooted in automation rather than asset creation. Although, Amazon’s Creative Studio also has plenty of tools available for brands interested in that particular facet of AI tech.”

Read AdExchanger. (May 11)

Related: “New research shows viewers don’t just want to watch your ads—they want to interact with them” (January 6) – Amazon

From tipsheet: Amazon is pushing Prime Video inventory toward a commerce optimization surface where creative, interactivity and purchase-stage messaging adapt using Amazon’s powerful shopping graph.

COMMERCE MEDIA

Bringing content to commerce with AI

Shopsense AI co-founder and President Bryan Quinn announced a new partnership with Amazon Publisher Services (APS) bringing his company’s content-to-commerce tech to APS partners’ connected TV screens. Read the release. (May 6)

MediaPost’s Laurie Sullivan covered the news, “[Shopsense AI’s] technology can ingest movie clips, photos, articles and other types of content, identify the images, with the images then transformed into a product listing ad to make commerce-enabled media more accessible.”

Later, Quinn discussed his company’s agentic technology:

“Shopsense runs agents to get the look and feel of the companies it works with — more than 1,000 retailers from apparel and home goods, to beauty, even auto parts, to determine how to create the ads, Quinn said.

‘The agent will determine what products to surface to consumers,’ he said. ‘Sometimes we find that the consumer really likes to engage with higher-price items. Maybe they are comfortable shopping and buying them.’”

Read: “Amazon Brings Content-To-Commerce Technology Into Publisher Services (May 11) – MediaPost

Related: The Trade Desk Taps Shopsense AI to Turn More of the Open Web Into Shoppable Ad Inventory (October 2025) – The Trade Desk

From tipsheet: The Content to commerce or “social shopping” category is about transforming media into optimizable commerce surfaces where products, offers and/or shopping prompts can be generated contextually from the content itself.

The themes of this announcement are reminiscent of Stratechery’s Ben Thompson’s interview of YouTube CEO Neal Mohan last September. Thompson wondered how long until YouTube would roll out scanning YouTube videos with AI and dynamically inserting relevant QR codes or links for shopping. Mohan demurred on a timetable but confirmed it was on the roadmap.

LLMs & CHATBOTS

Airbnb CEO on chatbots, designers

Brian Chesky, CEO of Airbnb, on TBPN yesterday:

  • “I do not think a chatbot is the right interface for travel or e-commerce.”
  • “If I were starting Airbnb today, I’d be vibe coding it in Claude Code.”
  • “I’m really worried that an entire generation of designers, artists and creative people are going to decide to sit out AI. And I think it’s the biggest opportunity for creative people in my lifetime.”

Watch: Brian Chesky, CEO, Airbnb interview (May 11) – TBPN on X

EVENTS

Prebid on buyer and seller agents

“Learn how Prebid is helping publishers drive more revenue with smarter, more automated ways to connect with buyers. We’ll share how seller and buyer agents are evolving, how feedback from buyers is shaping what comes next, and what it all means for your monetization strategy.”

Participants:

Register: “How Prebid Is Bridging Seller and Buyer Agents,” Thursday, May 14 – Prebid on Zoom

Related: The case for and against agentic media buying (May 11) – Digiday (subscription)

From tipsheet: Prebid develops and manages the open-source infrastructure and standards behind header bidding. Now, it’s increasingly positioning that infrastructure for agent-to-agent media buying and selling across the open web. Prebid took over stewardship and development of the Ad Context Protocol (AdCP) sell-side agent codebase in January.

MORE

  • IHeartMedia Revenue Rises on Demand for Digital, Podcast Advertising – The Wall Street Journal (subscription)
  • “Merchant Center for Agencies: Now Rolling Out Global” (May 11) – Ginny Marvin, Ads Product Liaison, Google on LinkedIn
  • Is Agentic Commerce An Oasis Or Mirage? (May 11) – AdExchanger
  • Podcast: Jim Payne of mobile sell-side ad platform CloudX joins investor Nihal Mehta (May 7) – Human Unicorn Podcast on Spotify
  • How AI is breaking advertising’s oldest business model – agencies (May 11) – exchange4media (India)