By Tim Peterson  •  February 4, 2026  •

This week we had NBCUniversal’s Ryan McConville on the Digiday Podcast to talk about the company’s test of using AI agents to sell ads against a live NFL game. It was a really insightful, technical conversation. And one thing McConville said has really stuck with me:

“Agentic AI and the ability to orchestrate these AI agents offer what I was calling premium automation,” said McConville, who serves as chief product officer and evp of ad products and solutions at NBCU.

As he explained, programmatic standards like the OpenRTB protocol have enabled computer programs to buy ads on behalf of humans. But the ad formats and inventory types purchased by those programmatic tools have been pretty simplistic, relatively speaking. By that I mean when compared to the complexity of a traditional TV ad buy, where the audience has to be forecasted ahead of time and there are all kinds of rules regarding which advertiser gets what ad slot in which program and what ad format to air there. 

“OpenRTB can streamline and automate the buying of certain things, but understanding bespoke ad formats, different sponsorship integrations – OpenRTB doesn’t really do that. But you can ask those questions of an [AI] agent and get those answers and start to automate some of that higher-end part of the market. And frankly TV is uniquely in that space,” said McConville.

In other words, AI agents can enable the automated buying and selling of traditional TV’s complex ad packages in a way that today’s programmatic advertising systems can’t support.

Honestly, those lines didn’t scream out to me in the moment. Not like they did the next day when I spoke with Frans Vermeulen, president of Swivel, a tech company that specializes in using agentic AI to automate ad operations.

Incorporating AI agents into TV advertising workflows “lets TV companies offer things that make them different from other media channels, which is these linked ad experiences: ‘I want to buy the first ad pod for tonight’s game for the first half. I want to be the last ad in every pod of the second half of the game. And then at the end of the game, I want to do some kind of takeover.’ That could never be traded through an RTB-type stack. That’s the big opportunity,” said Vermeulen.

If any of that TV-speak went over your head, here’s an analogy: Think of ad buying as like travel booking. 

• Historically there were human travel agents a person would call to book their trip, and that travel agent would contact airlines and hotels to arrange reservations. That’s akin to the traditional ad buying model. 
• Then online travel booking sites like Kayak came along to automate much of the process by plugging into airlines’ and hotels’ back-end systems. That’s akin to the programmatic ad buying model. 
• Now there’s the potential for AI agents to book travel so that a person could submit a prompt to ChatGPT along the lines of “I want to go to Tokyo sometime this summer. Determine when’s best for me to go given my schedule and local travel trends, and find me somewhere to stay that’s within walking distance of Shinjuku but not a tourist hub and ideally close to a train station.” And then ChatGPT books the arrangements like a human travel agent would have done in the past. That’s akin to AI agent-powered programmatic advertising.

But booking a flight and hotel is a pretty basic purchase. It’s like booking a banner ad or standard pre-roll campaign. TV advertising, though, can be a much more multi-faceted purchase package. It’s similar to booking not just a hotel room but a spa package and requesting a specific room in the hotel, plus early check-in and the hotel has a no-pets policy but can an exception be made with a deposit? 

Most travelers aren’t seeking out such an intricate travel package. But some are. And the latter are likely paying big bucks for those accommodations, which is why a hotel would tailor a reservation to meet the demand and reap the revenue. So too in TV advertising. And agentic advertising – via specifications like IAB Tech Lab’s new Agentic Direct spec – can make it so the TV ad market can finally automate ad packages that seemed incompatible with programmatic advertising. 

“Over the past 15 years, the TV industry has kind of been forced a little bit into trading through those kinds of programmatic protocols and systems,” said Vermeulen. “It has devalued what makes the TV industry – including streaming and all of its glorious flavors – unique from other media channels.”  

What we’ve heard

“I think they’re interchangeable.”

— YouTube creator Jesser on the difference between “creator” and “influencer”

Numbers to know

$110 million: How much money Disney lost during its distribution standoff with YouTube TV last year.

$30 billion: ESPN’s valuation after selling a 10% stake to the NFL.

27%: Amazon Prime Video’s share of global spend on sports rights among streamers.

44 million: Number of paid subscribers that Peacock had at the end of 2025.

What we’ve covered

Inside NBCUniversal’s test to use AI agents to sell ads against a live NFL game:

• NBCUniversal ran a proof-of-concept test with agency RPA, Newton Research and FreeWheel to have AI agents handle the buying and selling of an ad against a live NFL playoff game.
• NBCU’s Ryan McConville joined the Digiday Podcast to break down the agentic ad buying test.

Listen to the latest Digiday Podcast episode here.

Inside NBCU’s $3 million Peacock Super Bowl pitch:

• NBCUniversal asked advertisers to pay $3 million for a 30-second spot in Peacock’s Super Bowl stream. 
• It asked advertisers to match that spending on other Peacock inventory, making the minimum streaming entry point around $6 million.

Read more about Peacock’s Super Bowl ad sales here.

Why TikTok still faces an uphill battle in the U.S.:

• Despite completing its U.S. ownership deal, TikTok is facing backlash over changes to its privacy policy and rising data surveillance concerns.
• TikTok experienced a 195% spike in U.S. app uninstalls in the U.S. between Jan. 22 and 28.

Read more about TikTok’s U.S. challenges here.

Vita Coco rides out TikTok U.S. uncertainty with its safety nets built:

• TikTok accounts for around 70% of Vita Coco’s social spend.
• The brand hired creator Romeo to upload a sponsored video on the day of TikTok’s U.S. deal ratification, and that video racked up roughly 34.5 million views and drove Google searches for the brand to a four-year high.

Read more about TikTok marketing strategies here.

Brands are turning to predictive models for creator marketing:

• Brands like TheRealReal and SharkNinja are using predictive models and AI-powered datasets to identify effective creators and forecast campaign performance before spending a dollar.
• TheRealReal’s use of Fohr’s predictive tech yielded a 2.5x increase in views per dollar and a 440% lift in views for its holiday campaign.

Read more about predictive creator marketing here.

What we’re reading

Disney’s next CEO:

The company has named parks chief Josh D’Amaro as Bob Iger’s successor, according to CNBC.

Disney’s succession storyline:

The last time Bob Iger gave up the CEO didn’t turn out so well, raising the question of whether D’Amaro can succeed where Bob Chapek struggled, according to The New York Times.

SAG-AFTRA’s ‘Tilly tax’:

The union is considering requiring studios to pay royalties when using AI-generated actors, as AI increasingly threatens to replace human performers, according to Variety.

YouTube’s measurement issues:

Google sent a cease-and-desist to measurement bodies Barb and Kantar Media over their effort to include YouTube in traditional TV audience ratings, citing API violations, according to Tubefilter. This item has been updated to reflect that Google sent the letter to Kantar Media, not Kantar (the two are separate companies).

Paramount+’s short-form plans:

Paramount is exploring adding TikTok-style short-form videos on Paramount+ as Disney+ is also doing, according to Business Insider.