Agentic AI frameworks are growing

Agentic framework for ads

Ad Age’s Garett Sloane reported on Friday that Chalice AI has built an app that lives inside cloud-based data warehouse Snowflake. From there, marketers at pharma giant Bayer are creating a new custom algorithm for its programmatic ad strategy.

Sloane previewed the article on LinkedIn and observed, “It does speak to an agentic AI framework in ad tech, where apps will be a key method for brands to plug into ad tech services.”

Read more on LinkedIn. (November 7)

Read: “Inside Bayer’s ad tech and data strategy to find new customers and prepare for agentic AI” (November 7) – Ad Age (subscription)


LLMs & CHATBOTS

Why not ads in a chatbot?

On The Drum, reporter Kiri Masters disputes the idea that answer engines cannot have advertising because it will inhibit the consumer’s trust of — and experience with — the answer engine. She noted that ad tech companies like Thrad are taking cues from the new Ad Context Protocol (AdCP), and developing for “conversational advertising.”

Previewing her article on LinkedIn, Masters described the opportunity for AI chatbot companies succinctly: “Advertising represents trillions in addressable market opportunity—it’s not quite ‘free money,’ but it’s pretty damn close.”

Read more in The Drum. And, read Ms. Masters on LinkedIn. (November 7)

From tipsheet: Some AI company with an LLM supporting a chatbot, and looking for a revenue model, will take the ads “leap.” From here, it seems too easy. Also, a first mover in “ads in a chatbot” could unlock benefits that no one knows today.

How about using audience data for chatbot targeting, too? Just brainstorming…


SELL-SIDE

AI impact on search traffic two-fold

Ad tech firm Teads is still working through its merger of Outbrain and Teads which completed in February. Teads Q3 2025 earnings were released last week and were slightly below analysts’ expectation. Read the release. (November 6)

On the earnings call with Wall Street, CEO David Kostman (former Outbrain CEO) suggested that changes to the way consumers are searching – using answer engines versus traditional search – may be partially to blame.

DAVID KOSTMAN: “We continue to experience declining page views on premium publishers, partly due to increased adoption of AI summaries and volatility in our programmatic supply. However, this has been partially offset by ongoing RPM improvements and by actions taken by publishers to increase engagement of the audiences, particularly on their applications.”

Needham & Co. analyst Laura Martin picked up on the comment and asked Teads CFO Jason Kiviat about the AI impact.

LAURA MARTIN: “One of the things you said at the top of your comments was that you are seeing you’re the first actually ad tech company that’s reported that says they’re seeing a diminution in traffic. Magnite said they’re hitting record traffic levels even excluding bots. So I’m curious about that. Do you think that’s because your content is primarily news? And that also sounds structural. So can you talk about this, the traffic demise that you’re seeing that at least other CEOs are not admitting to? So I’m interested in what you’re seeing on the traffic side.”

JASON KIVIAT: “On the traditional publisher side, when we look at the sort of list of premium publishers, we saw around between 10-15% decline in page views. I mean, these are the numbers we are seeing. I think it’s very consistent with everything you’re reading out there.

So if everyone is saying that there’s no decline in publisher page views, I suggest you do a ChatGPT and you’ll see those numbers. What we see, I think it’s a little bit softer on in-app traffic. In-app traffic is about 30% of those publishers’ traffic, and there we see still some decline in the page views lower than that. So single-digit on the in-app, and on the web around 10 to 15%. That’s what we see on a certain segment of publishers that I believe is representative.”

Read the earnings transcript. (November 6)

Related: For Publishers, AI Gives Monetizable Data Insight But Takes Away Traffic (November 7) – AdExchanger

From tipsheet: Interesting to hear Teads’ statement that revenue per thousand (RPM) page views has gone up for publishers in spite of decreased overall page views. Chatbots appear to be scrubbing the open web of less valuable traffic.


BRANDS

Forrester on open web display budget

Forrester analysts are predicting that some brands are ready to cut their display ad budgets on the open web by up to 30% due to AI search.

Digiday quoted Forrester Research ad tech senior analyst Evelyn Mitchell-Wolf: “We expect that fewer consumers will be landing on the web pages where most of that display inventory has historically existed, and so there will just be fewer monetizable audiences there to begin with.”

But, it’s more reallocation than cutting, as Digiday noted. Forrester sees the spend going to CTV and paid social. Read more. (November 6)

From tipsheet: if you’re a marketer and heard the Teads earnings call (see “AI impact” summary), you may not have the scale you want, but the open web’s improved quality may be enticing. In the long run, does this speak to open web viability? Proof will be in the “outcome pudding.”

Related: “AI is rewriting visibility in the zero-click search era” (November 4) – MarTech


BRANDS

Use case: Contextualizing human feelings

Speaking at last week’s ESOMAR (a community for data, research, and insights) conference in Washington D.C., a marketing analytics executive at beverage company AB InBev shared how his company “used AI-powered language analysis and psychological theory to zero in on the best creative approach.” The AI tech was provided by Brand Genetics for a campaign promoting a non-alcoholic beverage, Corona Cero, in 2024. See the video ad.

Sebastian Schuliaquer, global director of premiumization insights and foresight at AB InBev, said, “AI was helping us contextualize what people were feeling. Two years ago, there were a lot of AI-generated campaigns, and it was very scary, because there was beautiful stories with no soul. We needed to have that human element, and we found it in this space.”

Read more on Marketing Dive. (November 7)


AGENCIES

Agencies training for AI transition

  • Agency GSD&M gamifies its AI upskilling efforts (November 6) – Digiday (subscription)
  • AI is reshaping entry-level roles—how agencies and schools are retooling (November 7) – Ad Age

CONSULTING

AI workflows for creative, media

On Friday, Boston Consulting Group (BCG) trumpeted the results of a new survey of senior marketers about the use of AI in marketing.

A trio of authors, including former Tremor Video CEO Lauren Wiener, interpreted the survey and found “low-hanging fruit” for marketers in creative and media writing, “CMOs should focus on redesigning end-to-end workflows where AI disruption is likely to be greatest and most rapid.”

Other BCG takeaways:

  • “Although AI can automate basic marketing processes, CMOs who focus exclusively on productivity gains are missing the bigger picture.”
  • “Marketing leaders who fully embrace AI can transform their entire operating model in ways that were previously impossible.”
  • “The challenge for CMOs is to sell their marketing vision to the C-suite and to their own teams while forging new partnerships with outside agencies.”

Read more. (November 7)

Related: “The state of AI in 2025: Agents, innovation, and transformation” (November 5) – McKinsey & Co.


PLATFORMS

Opinion: Agent recommendations

On his personal blog, consultant and university professor Shelly Palmer weighed in on Amazon taking issue with Perplexity’s Comet browser and its agentic browsing of Amazon’s websites.

Mr. Palmer believes that Amazon’s consternation is understandable considering that Amazon’s $60 billion sponsored ad revenue may be at risk. And then, he zoomed out:

“Agentic browsers won’t just search Amazon. The more you use a shopping agent, the more it learns your buying habits. It compares prices across dozens of retailers, factors in shipping costs and return policies, and finds the best overall deal. This will either level the playing field for all online retailers or favor those who pay. We can all guess where this is going. Retailers with the biggest budgets will pay for placement in agent recommendations, just like they do in Google search results.”

Read more. (November 9)


ANALYTICS

AI rewriting customer lifetime value

Daniel McCarthy, a co-founder of predictive analytics firm Theta, shared his latest research on “what AI is doing to customer lifetime value and discovery” after having spoken at a recent Google Data & Analytics Summit

He explained:

“What hit me while building my keynote is that the customer acquisition chapter of the ‘customer lifetime value (CLV) textbook’ is being completely rewritten right now. Discovery, the entire front-end of the customer adoption funnel, is completely different from how it was even 18 months ago. A massive chunk of it is in the process of migrating to LLM’s.”

Read what he means on LinkedIn. (November 9)

h/t Eric Seufert


PODCASTS

Talking AI

  • “The Scape of Things (with guest Terence Kawaja)” (November 6) – The Brand New podcast with Marisa Thalberg and Stephen Wolfe-Pereira on Apple Podcasts app

  • “There Will Be Blood (ft. Joe Zawadzki, GP, Aperiam Ventures)” (November 7) – AdTech AdTalk podcast on YouTube


SELL-SIDE

Mozilla AI content marketplace

Sell-side advocate Matthew Scott Goldstein believes Mozilla is on to something with its new AI content marketplace. He wrote on LinkedIn Friday, “Data set owners have the flexibility to set their own prices for usage or impose specific limits on who can access the data, under what terms, and for what purposes.” Read more on LinkedIn.

See more: Mozilla Data Collective


TECH

Meta -and AI vs. fraud

A report by Reuters on Thursday alleged that Meta internal documents knowingly projected approximately 10% of company’s 2024 revenue ($16 billion) would be “from running advertising for scams and banned goods.” Furthermore, Reuters said a 2023 planning document revealed that “everyone who worked on the team handling advertiser concerns about brand-rights issues had been laid off. The company was also devoting resources so heavily to virtual reality and AI that safety staffers were ordered to restrict their use of Meta’s computing resources. They were instructed merely to ‘keep the lights on.’” Read the report. (November 6)

On his newsletter, Platformocracy, former-Googler-now-author Jonathan Bellack was getting traction in social media circles when he took issue with not only the findings but Meta’s public relations response. Read that one. (November 7)


Google AI Max critique

Retail platform Smarter Commerce reviewed its latest research of client accounts using AI Max, Google’s AI-enabled search marketing platform.

In addition to limited adoption (at least at the account level), Smarter Commerce saw mixed results with AI Max. Specific concerns included:

“• AI Max is pushed toward ALL advertisers, even those with mature Search campaigns
• AI Max often targets traffic that was purposely not targeted before (e.g. competitor terms)
• AI Max technology is redundant with DSA and PMax, complicating account hygiene
• AI Max is marketed as the gateway toward AI placements (e.g. AI Mode), even though it affects the whole campaign, and these placements can be served with Broad Match.”

Read more on LinkedIn. (November 6)

From tipsheet: Still feels early days for some/all AI-enabled automated platforms?


MORE

  • A.I. Sweeps Through Newsrooms, but Is It a Journalist or a Tool? (November 7) – The New York Times (subscription)
  • firmly Unveils Unified “Buy Now” Platform to Power the Future of Agentic Commerce (November 5) – press release
  • When It Comes to AI, Adapt or Else, Says [Book Publisher] (November 6) – Publishers’ Weekly
  • Discover how Kokai’s AI powers measurable business outcomes (November 6) – The Trade Desk
  • Why tech giants are offering premium AI tools to millions of Indians for free (November 7) – BBC