AI creative making the case for AI

John Ebbert

Making the AI case

Google is being careful about how it uses AI in its own advertisements. A new ad with an animated, turkey plush toy using AI is a proof point. (See YouTube)

The Wall Street Journal’s Patrick Coffee took note and previewed his latest article on LinkedIn: “Google, unlike other marketers ‘drunk on AI’ and rushing to brag about their hyper-realistic but totally fake videos, isn’t quite ready to go there.”

From Coffee’s WSJ article:

“The broader ‘Just Ask Google’ campaign, which started earlier this year, is designed to position Google’s AI search features as part of its product suite, particularly among people who may be apprehensive about the technology or even scared, said Marvin Chow, vice president of marketing for Google AI and DeepMind, its AI research division.”

Read the article. (October 31)

From tipsheet: A brand ad with an AI-generated plush toy will be one way among many that Google will try to ring-fence its AI-enabled future.

OpenAI chose no use of AI-generated video whatsoever in its recent brand campaign.

A robust government relations team will remain critical for all of Big Tech including the big, private AI companies.

LLMS & CHATBOTS

Developments

AGENCIES

WPP and optimizing the black box

eMarketer’s Senior Director of Content Jeremy Goldman empathized with WPP’s new CEO Cindy Rose regarding WPP’s previous lack of urgency and inability to move at “AI speed.” Mr. Goldman suggested on LinkedIn:

“For marketers, this is a turning point. The value of scale is fading; the value of speed, insight, and specialization is soaring. Holding companies were built for reach, not responsiveness, and the market’s now voting with its budgets.

The good news: WPP still has incredible creative firepower and blue-chip clients. The bad news: those assets mean less if execution lags. As Rose begins her reset, WPP’s challenge isn’t just about tech adoption; it’s about cultural velocity.”

Read the thread on LinkedIn. (October 31)

Alan Morrissey, a Data Strategy Partner at WPP Media’s Choreograph, took issue with Goldman’s argument in the comments:

“Problem is media is not a one platform game, the walled gardens can launch their own tools, the power & value comes form being able to plan & activate accordingly across multiple platforms…….yes they form a large part of activation (rightly or wrongly), but ask any media planner about the complexity of a full-funnel campaign, it ain’t simple and AI won’t fix that…..WPP Open Pro allows for access to 3PD data for insights & activation across multiple platforms….”

More: WPP’s Open Pro AI suite already faces competition from Google and Canva (October 31) – Digiday

From tipsheet: Marketers, agencies and tech companies alike are going to target the “cross-black box” optimization future. What’s possible, and/or what compromises will be involved, as black boxes get even more tightly wound, are hard to predict.

SELL-SIDE

Microsoft’s Content Marketplace and Gannett

Publishing megacorp Gannett said during its Q3 earnings announcement last week that it would be a publishing partner in Microsoft’s new Publisher Content Marketplace (PCM).

Andrew Steinberg, a senior biz dev executive at Gannett publicly thanked on LinkedIn six members of the AI publishing marketplace team at Microsoft including Nikhil Kolar, Devna Shukla, Kathy Henwood, Geraldine Russell, Sayantan Roy, and Kavitha Lingamoorthy.

Read: Gannett Announces Third Quarter 2025 Results, New AI Licensing Deal with Microsoft & Updated Business Outlook (October 30) – Gannett

From tipsheet: There are no details on the specifics of the licensing agreement. As I noted in September, “This appears to be the next step of a plan that Microsoft started last year when it began to compensate publishers for AI usage. Flat fee, AI licensing deals for AI crawling privileges are numerous with usage-based deals still to come.:

TECH

A buy-side FAQ on AdCP

Joseph Hirsch, CEO of AI ad ops automation platform Swivel, announced that his company is helping clients get “set up to sell agentically” using Magnite’s SpringServe adserver today as well as future integrations planned with FreeWheel, IAS’ Publica, Google Ad Manager and Kevel.

He also shared an FAQ he’s compiled for Ad Context Protocol (AdCP). On Friday, he offered a selection of his sell-side questions and answers on Linkedin:

“What are the economics of an agentic transaction?

Sellers can clear through Swivel or directly with the buyer for a single digit transaction rate.

What can I expose to seller agents?

Sellers can expose a range of things that generally wouldn’t flow through the bid stream. This includes first party data like ACR data or transaction history for an RMN. Show level data, content, and context, as well as a full range of traditional and custom inventory. Because a seller agent doesn’t store or process data, and only uses targeting inside a seller ad platform, traditional issues like data leakage, or sharing sensitive information to a DSP are moot. Data and targeting information never leave the confines of a seller’s first party ad platform.

What is an ad product?

An ad product is either a static or dynamic combination of inventory, audience, content/context, and price…”

Read more Qs and As on LinkedIn. (October 31)

From tipsheet: Mr. Hirsch promised a buy-side-focused AdCP FAQ shortly.

Ads are the future of AI search

  • Google’s Robby Stein: Ads aren’t going away in AI search (October 31) – Search Engine Land

  • Google VP: SEO and AI search optimization have ‘a lot of overlap’ (October 31) – Search Engine Land

RETAIL MEDIA

SKU data as the fuel for AI

Former Turn demand-side platform executive David Simon, who is now president of in-store advertising company Vibenomics, has sharp views on what retailers should be doing regarding AI. It begins with their own data.

Simon shared on LinkedIn on Friday:

“There is a ton of noise in the retail media market about AI enablement of smarter ad decisioning but frankly, I’m calling BS. AI only works well when there is a direct connection between ad dollars spent and products sold. (Look at mobile advertising for the broader media industry’s best example of this working well.)

(…) Retailers need to think about their SKU-level sales data as the fuel for an AI machine. The faster we get there, the faster the RMNs will scale revenue…when they can show that a dollar in equals two dollars out.”

He pointed to an October Path-to-Purchase article titled, “3 Ways AI Will Bring Category Management Into a New Era.” Read it here. (October 28)

From tipsheet: The more transparency (including via data) for AI, the better the outcomes. Except… AI operates in a black box.

FINANCIALS

What ad market slowdown?

Online advertising is buoying Big Tech earnings results, according to CNBC. Jasmine Enberg, a former eMarketer analyst, said: “I think the digital ad market is strong. I think this economic instability and volatility is kind of priced in for a lot of people at this point; sort of seems to be the status quo.” More to come?

Speaking on an earnings call with Wall Street on Thursday, Amazon CEO Andy Jassy said he has high expectations for Amazon DSP: “You look at some of the partnerships that we’ve done, the Roku partnership gives us the largest connected TV footprint in the U.S. And you layer on top of that what we’ve recently done in providing our DSP customers the opportunity to integrate with the ad inventory in Netflix and Spotify and SiriusXM, it’s powerful.”

Read more on CNBC. (November 1)

More: What investors learned from tech earnings, in charts (November 1) – The Wall Street Journal (subscription)

SEARCH

Answer engines offer views, not clicks

On LinkedIn, Chief Information Officer Alex Sukennik of Semrush, a search engine marketing platform firm, shared a brief demo of his company’s new “visibility” product which now spans answer engines, Semrush One.

Note the pitch:

“Search is changing exponentially. From traditional search to LLMs, buyers move between Google, ChatGPT, Gemini, Perplexity, and more, often without ever clicking a link.”

Read a bit more on LinkedIn. (October 31)

And that appears to be a key difference in the world of answer engines and LLMs. With SEO, the ultimate goal is a click by virtue of a listing high up the search engine result (SERP) page. With AEO (answer engine optimization), the ultimate goal is a view.

Reddit, the AI publisher ‘darling’, said as much last week: “Chatbots are not a traffic driver today.”

From tipsheet: Understanding an answer engine view’s impact will bring new life to the view-through conversion. Also, answer engine visibility metrics will provide another data feed for incrementality attribution models.

FINANCIALS

Earnings reports this week

AI’s impact —particularly on the open web— will be on display once again in this week’s earning reports.

From tipsheet: Any crumb of information about its partnership with OpenAI is something to watch on the Shopify earnings call. How OpenAI makes money remains the media’s foremost fascination.

TECH

Performance and The Trade Desk

On the Marketecture podcast, LUMA Partners Terence Kawaja shared his ‘take’ on the M&A landscape for advertising technology today – and on The Trade Desk, in particular.

He told Marketecture founder Ari Paparo at “Marketecture Live”:

[31:00] “I believe AI is going to accelerate the whole move to outcomes. I also think we need to reframe what we mean by outcomes. We do not necessarily mean, you know, direct response, bottom of the funnel, final step before conversion. No. (…) Outcomes is a full funnel phenomena… And I think that [The Trade Desk is] missing a big opportunity in performance.”

He suggested to Mr. Paparo that The Trade Desk should buy Moloco.

Podcast: “LUMA Partners’ Terry Kawaja lets loose on the Marketecture Live stage” (October 31) – Marketecture podcast

Moloco was last valued at $2 billion during its 2023 funding round. The Trade Desk’s market cap is currently $25.4 billion.

From tipsheet: Interesting discussion between Messers. Kawaja and Paparo about The Trade Desk creating a performance engine using the open web.

TECH

Egress and cloud competition

On their AdTech AdTalk podcast, Gareth Glaser and Adam Heimlich offered “teacher mode” in the context of AWS’ recent RTB Fabric announcement —which made it more inexpensive to do ad tech business on the AWS platform via containerization.

But, the two ad tech executives believe everyone needs to understand “egress” fees to truly understand Fabric’s impact.

And so, a snippet from the podcast included Mr. Glaser defining “egress”:

“Basically, when data leaves a system —when they need to send data to somewhere else, it consumes something called ‘bandwidth.’ And bandwidth is the pipeline that exists between those two systems that can transmit a certain amount of data at one time.

So, most of the big cloud providers —Amazon included— charge a lot of money for egress to other systems. If you’re sending data from inside Amazon to somebody outside of Amazon, you pay a bandwidth rate. That’s actually variable by geo. This tends to be the largest cost center in operating real-time bidding (RTB) infrastructure.”

Hear the clip on Linkedin. (November 2)

AWS said the impact of RTB Fabric for customers is “up to 80%” reduction in network, or egress, fees.

Podcast: “Meet the Moment (featuring Stephanie Layser, AWS)” (66 minutes) – AdTech AdTalk podcast (October 31)

From tipsheet: Ultimately, Amazon Web Services is looking to facilitate more business within their platform. Also, in an AI world, data will become even more voluminous. By offering a product (RTB Fabric) that offers easier and less expensive integration to partners via AWS, there’s less of a reason for ad tech companies to go elsewhere for cloud services.

MORE

  • Understanding The Trade Desk’s Competitive Dynamics (October 22) – Matthew Blake
  • With Canva Grow, is SaaS about to eat the performance creative industry? – Henry Innis
  • Eskimi partners with Index Exchange: Scaling AI-driven contextual activation (October 23) – Eskimi
  • Divulging Brand Secrets Rises With AI Agent Concerns (October 30) – MediaPost