Automated buying is evolving on The Trade Desk’s AI-enabled Kokai platform.
Digiday reported that new automated “Trading Modes” are continuing to move forward in closed beta with select clients. For The Trade Desk, the product strategy reflects a response to today’s digital ad market where advertisers want both control and automation options.
Originally announced last September, the modes break out as follows:
- Control Mode: Clients can execute traditional, granular buying.
- Performance Mode: AI handles execution including optimization, bidding, and allocation.
In addition, a consolidated, campaign management fee structure is introduced which bundles media, data, and tech fees into a single price.
Read:
- “The Trade Desk is changing how advertisers buy — and what they can see” (April 6) – Digiday (subscription)
- Trading Modes (Closed Beta) – The Trade Desk
- The Trade Desk Announces Major Overhaul of Digital Advertising Data Marketplace (September 2025) – The Trade Desk
Related: CNN builds in-house agent infrastructure as it prepares for AI-driven media trading (April 6) – Digiday
From tipsheet: Programmatic is heading towards less hands-on execution, more outcome-driven buying.
LLMs & CHATBOTS
Developments
- Industrial policy for the Intelligence Age (April 6) – OpenAI
- Meta to open source versions of its next AI models (April 6) – Axios
- Anthropic expands partnership with Google and Broadcom for multiple gigawatts of next-generation compute (April 6) – Anthropic
MARKETING
The in-housing trend: SEO and GEO
Yesterday on LinkedIn, Pfizer media lead Josh Palau announced that his team had in-housed its search engine optimization (SEO) and generative engine optimization (GEO) needs in-house.
He wrote:
“We have officially transitioned SEO & AI Discoverability to the Pfizer In-House team!
What I love most about this? It wasn’t even part of the original plan. This was born out of the vision of Sarah Clayton and Sari Elzweig, who saw the GEO opportunity coming, and recognized how it’s reshaping how our brands are discovered, understood, and trusted.”
Read more on LinkedIn. (April 6)
RETAIL MEDIA
Dunnhumby seeks agentic commerce intelligence
On James Avery’s “Unlocking Retail Media” podcast, Michael Schuh, Global Head of Media at dunnhumby, discusses the future of data-driven retail media including his views on AI’s impact.
Dunnhumby is the retail data science company backed by UK retailer Tesco.
#1 – In-store
Schuh is a huge believer in the in-store “moat” for retailers as well as the media opportunity that can be unlocked.
He said: “I am really excited about the focus in-store. (…) 80 to 90 percent of the transactions are still happening in-store. It’s an environment where I think the customer has to be even more at the forefront of the decisions that are made.”
#2 – Agentic commerce
Regarding “agentic commerce” and whether autonomous agents will buy on behalf of consumers and potentially disterintermediating media revenues from retailers, Mr. Schuh is not dismissive of the threat it poses retail media.
Nevertheless, he sees an agentic opportunity in intelligence, saying: “A huge part of what we see as a potential and what we’re investing in are the AI capabilities that can sit on top of data and insights engine that help proactively show insights…. [For example,] ‘Here’s a suggested campaign for you to actually run through this retailer to drive the best performance’ — and maybe you got some ability to automate parts of that.”
Listen: “From Unlocking Retail Media: The Power of Customer Data Science: Lessons from the Front Lines with Michael Schuh” (March 25) – Unlocking Retail Media on Apple Podcasts app
From tipsheet: In theory, sophisticated mega-retailers have a potential advantage in the next big leg of AI transformation: the physical world, in-store and/or out-of-home.
This could be AI’s biggest unlock in advertising given the 80-90% of in-store transactions referenced by Schuh.
How about serving media in-store via conversational AI shopping assistants which gradually become the consumer’s new best friend?
RETAIL MEDIA
Test now: Conversational ads and the retailer
Growth marketer Todd Piechowski of Envision Horizons laid out his ChatGPT ads plan for retail media strategists in a post on his aptly-named “Instant Checkout” blog.
(The blog promotes his eBook by the same name and is no relation — well, loosely — to OpenAI’s mothballed Instant Checkout.)
According to Piechowski, retailers must understand how consumers’ most important, new discovery mechanism works with ads — i.e. the conversational chatbot.
He advocates for use of still-unannounced access to ChatGPT ads that — he says — is coming April 14:
- “Don’t let FOMO drive a $200K bet. Self-serve access is coming this month. Wait for it. Test at a spend level where broken attribution doesn’t matter.”
- “If you’re already running retail media on Amazon or Walmart through a platform like Pacvue, pay attention to what they announce April 14. Managing ChatGPT from the same dashboard you use for Amazon changes the operational cost of testing dramatically.”
- “Google AI Mode is the more measurable bet right now for most brands. If you’re running Performance Max, you’re already there. Check your AI Mode placement data before you chase a new channel.”
- “Watch Perplexity. If the ad-free model wins consumer trust and starts pulling market share, that changes the math for every AI platform running ads. The ‘Google search model applied to AI’ assumption isn’t guaranteed.”
- “Regardless of ads, get your product data agent-ready. Every one of these channels — ChatGPT, Gemini, Rufus, Perplexity’s Merchant Program — runs on structured product data. The brands that show up cleanly in AI recommendations, sponsored or not, are the ones with complete, consistent, real-time catalog feeds. That’s the work that compounds no matter which channel wins.”
Piechowski is skeptical that ads in ChatGPT will be performant for retailers and wonders if generative engine optimization (GEO) may be the optimal way forward.
Read: “ChatGPT Just Became a Retail Media Channel. Nobody’s Ready.” (April 5) – Instant Checkout blog
Related: “How AI Helps Scale Qualitative Customer Research” (April 6) – Stefano Puntoni, Jeremy Korst and Olivier Toubia in Harvard Business Review
From tipsheet: Does he know something? We’ll see if the ChatGPT ad floodgates open on Tuesday, April 14.
Not to be underestimated in this extended test phase is that OpenAI is working hard to serve ads that perform. Is it the 17th prompt appropriate for an ad in a toothpaste discovery conversation? It is not simply: “Bring your ads here, retailer, and we’ll put it in the conversation with some contextual category targeting.”
Then the question becomes what is the true scale of the opportunity?
The opportunity could be massive.
SELL-SIDE
Keyword blocking inhibits context
Media buyers are still blocking with keywords in spite of better contextual analysis enabled by AI.
In AdExchanger, IAS Chief Product Officer Srishti Gupta reviews a recent test across Reuters websites which compared keyword blocking to IAS contextual blocking technology. The test once again highlighted a recurring problem for news publishers who must deal with a risk-averse buy-side looking to prevent an ugly adjacency for their media at all costs — including reach.
Machine learning + semantic technology to the rescue.
AdExchanger’s Anthony Vargas reported:
“Gupta said, IAS prefers that advertisers use Context Control Targeting over keyword blocking. She also pointed to a solution called Quality Connect, which gives publishers visibility into how their content gets blocked by buyers so they can diagnose and correct any issues with overblocking.
Despite the availability of smarter tools, simple keyword blocking persists as a strategy. And both publishers and advertisers are paying the price.”
Related: Opinion – “AI Has Already Decided: First-Party Data Will Define Advertising’s Agentic Era” (April 6) – Alvaro Palacios, CSO, Newsweek on AdExchanger
From tipsheet: Adjacency concerns should lessen as AI enables better performance. Put another way, “Drop the keyword targeting because it’s inhibiting your ability to maximize your desired outcomes.”
AGENCIES
Sorrell renews AI agency mantra
Martin Sorrell, chairman of agency holding company S4 Capital, renewed his mantra on agency services in the age of AI.
In a Fortune article published yesterday, he began with outcomes:
“Audit how you charge. If your revenue depends on how long tasks take your team, you’re exposed. Move toward pricing based on what you deliver, not how many hours it consumed. Maybe you’re the founder who bills 40 hours for a content package that AI helps you produce in 10. That gap is your vulnerability and your opportunity. Close it before your clients do the maths.”
And then… personalization at scale:
“The second step is using AI to produce huge volumes of personalised assets. Where you once created one campaign and hoped it landed, now you produce dozens of variations tested against specific audiences. Sorrell sees this as an expansion of opportunity. More content, more formats, more touchpoints. The business model shifts again toward output pricing because the volume of work explodes.”
Read: “How agencies survive AI, according to the man who built advertising’s biggest empire” (April 6) – Fortune (subscription)
From tipsheet: Charge for outcomes and personalize at scale. Easy enough.
Now… what’s an outcome?
Or more appropriately, what’s an acceptable outcome for both the services group and the marketer?
PLATFORMS
Netflix, RecSys and personalization
New research titled, “Netflix Artwork Personalization via LLM Post-training” (January 6) from Netflix engineers inspired the latest musings of analyst Eric Seufert on his blog, Mobile Dev Memo.
Exploring the intersection of RecSys and large language models (LLMs), Seufert unpacks how LLM’s ability to understand context is unlocking a new level of personalization for recommendation models.
Netflix research activity is “signal” as Seufert says, “Given its institutional dedication to RecSys and the business value it attributes to it, any successful applications of novel personalization and recommendation techniques by Netflix are worth paying attention to.”
Read: “RecSys and chill” (April 6) – Mobile Dev Memo (subscription)
Related: Criteo CEO Michael Komasinski on RecSys and LLM usage (April 2) – tipsheet
SEARCH
For SEO, humans make better content than AI
The results from survey data of over 200 search engine optimization (SEO) professionals and tens of thousands of blog pages showed the effects of human versus AI-generated content.
TL;DR – The originality of human content won by a large margin.
SEO/AEO research firm Semrush (which was acquired by Adobe in November) observed:
- “72% of SEOs say AI content ranks at least as well as human-written content. But our analysis of 42,000 blog pages shows position 1 results are 8x more likely to be human-written.”
- “70% of SEO teams cite speed as the top benefit of using AI, but only 19% say it improves content quality.”
- “AI use is concentrated in text-based tasks but drops sharply for multimedia and localization, where the tools are more specialized and the output is harder to evaluate.”
Read: “Does AI content rank well in search? [Survey + Data study]” (April 1) – Semrush
Related: “An AI Storytelling Startup is On Pace to Generate $100 Million in Annual Sales” (April 6) – The Information (subscription)
From tipsheet: AI is not as powerful as SEOs think?
AGENCIES
Stagwell’s AI Search platform
Yesterday, agency holding company CEO Mark Penn touted Stagwell’s new generative engine optimization (GEO) — or AI search — solution originally announced in early March.
“Last week Google expanded Search Live across 200 countries, meaning just about everyone in the world can now interact with its AI mode using voice and camera. For marketers, understanding how their brands show up in AI environments and LLMs is becoming more and more essential, which is why we’ve launched Stagwell Search+, marketing’s first agentic platform to win AI search.”
Read more — and see his video — on LinkedIn. (April 6)
Promotion for the new AI search visibility platform includes a use case involving Google’s AI Max and paid search ads. Stagwell claims “+34% increase in revenue from AI-powered search transformation with a global technology client using AI Max.”
More: Stagwell Launches Stagwell Search+: The Industry’s First Agentic Platform To Win AI Search (March 2) – Stagwell
From tipsheet: If LLMs learn from the web, and ads fund and shape the web, then paid media may quietly influence what AI surfaces.
Ads are content.
MORE
Now hiring
- Long-time Aol/Verizon/Yahoo DSP product executive Daniel Spring joins The Trade Desk as “Senior Staff Product Manager, RTB and Inventory” (April 6) – Daniel Spring on LinkedIn
- James Goldstein joins Adgentic as Chief Revenue Officer (April 6) – Adgentic founder Ben Brenner on LinkedIn
- Jaclyn Petrovich joins Bluefish as Marketing Lead – Jacklyn Petrovich on LinkedIn
MORE
- Brands Adopt ‘No AI’ Disclaimers to Stand Out Amid the Slop (April 6) – The Wall Street Journal (subscription)
- Podcast: “Unity copies the AppLovin playbook with Vector” (April 6) – pocketgamer[dot]biz
- Opinion: “Brands Beloved by People Risk Being Invisible to AI” – Craig Elimeliah, chief creative officer, Code and Theory in Adweek
- Marketing firm’s AI bet: fewer repetitive tasks, more creative work (April 6) – Jacksonville Business Journal (subscription)
- How Magnite’s Live Scheduler Powers and Simplifies Scripps’ New Sports Channel (April 6) – Magnite