Back in November, Amazon sent a cease-and-desist letter to provocative AI company Perplexity and requested that it stop crawling Amazon’s eCommerce websites via agents tied to Perplexity’s Comet browser.
From the letter:
“Perplexity AI, Inc. has persistently refused to operate transparently with respect to its agentic AI product in the Amazon Store. Perplexity must immediately cease using, enabling, or deploying Comet’s artificial intelligence (“AI”) agents or any other means to covertly intrude into Amazon’s e-commerce websites (together, the “Amazon Store”) in violation of the Computer Fraud and Abuse Act (18 U.S.C. § 1030) and the Comprehensive Computer Data Access and Fraud Act (Cal. Penal Code § 502).”
Perplexity didn’t stop and argued it was being unfairly bullied.
But yesterday, the court sided with Amazon and ordered Perplexity to stop its crawling of Amazon’s shopping sites.
Read: “Judge blocks Perplexity’s AI bot from shopping on Amazon in early test of agentic commerce” (March 10) – GeekWire
Related: Perplexity accused by Cloudflare of stealth (August 2025) – tipsheet
From tipsheet: The seeds of the disagreement are: Perplexity inserting its agent between the user and Amazon, the valuable transaction data involved, and ultimately the threat to Amazon’s $60-70 billion advertising business which the data enables.
Whether Perplexity appeals, or what the next steps are, remains to be seen. But let’s not forget the tangled tech ecosystem — Perplexity is likely still an Amazon customer. When Amazon invested in AI company Anthropic in November 2024, the announcement included this tidbit from Anthropic regarding one of its clients: “Perplexity, an AI-powered search engine, delivers more accurate responses at twice the speed by using Claude in Amazon Bedrock.”
So like many tech brouhahas, one has to wonder if a business partnership is the end game. Perhaps an agentic commerce agreement which facilitates the interests of both companies — even if it’s just a test. Amazon now has more leverage to dictate terms.
LLMs & CHATBOTS
Developments
- Meta hires duo behind Moltbook (March 10) – Axios
- Amazon launches Health AI agent on Amazon website and app with free 24/7 access to virtual care for Prime members (March 10) – Amazon
- Health Check: How People Use Copilot for Health – Microsoft
PROTOCOLS
The value exchange protocol
Bearing the fruits of one of its Working Groups, IAB Tech Lab released for public comment the latest iteration of the Content Monetization Protocol (CoMP) yesterday. CoMP aims to enable a value exchange between content owners and AI companies.
All comments are due by April 9.
Tech Lab CEO Anthony Katsur teed up CoMP’s significance on LinkedIn:
“Today, the IAB Tech Lab is proud to introduce the Content Monetization Protocols (CoMP) framework, a new technical standard designed to support commercial agreements between large language models and publishers when using their content. (…)
This marks a key step toward a balanced AI ecosystem that supports a long-term, economically viable content marketplace, allowing media companies to establish greater control over how their content is accessed, supports informed business negotiations, and establishes clear sources of truth for both brands and media companies when consumers query LLMs.”
Read more. (March 10)
In the press release, Jennifer Bas, Chief of Staff at ad tech firm Mobian said, “CoMP creates a mechanism for that quality and trust to be valued in the AI supply chain.”
VP Global Data Standards Achim Schlosser added on behalf of publishing company Bertelsmann: “As an early contributor, we believe scalable, robust compensation frameworks — alongside visibility and attribution for content usage — are essential to sustaining high-quality journalism and premium content in the AI era.”
Read: IAB Tech Lab Announces CoMP Framework to Ensure LLMs Have Commercial Agreements with Publishers Before Content Crawling (March 10) – IAB Tech Lab
More:
- Content Metadata Marketplace Supply Specification (MVP) – GitHub
- Announcing “Content Monetization Protocols (CoMP) for AI” Working Group (2025) – IAB Tech Lab
From tipsheet #1: This is one of several efforts looking to assist with the AI-publisher value exchange. Firms such as TollBit already provide a “toll booth” solution for scraping with some success. Cloudflare has made announcements regarding finding solutions (1, 2) as part of their content delivery network offering. CEO Matthew Prince has been ‘banging the drum’ about AI crawlers since last June.
Will it be a pay per crawl model, pay per inference or? Also – how programmatic can it get? So much to discover.
From tipsheet #2: The well-known large language model (LLM) providers (OpenAI, Google, Anthropic and so on) would seem to be an important part of the discussion and are missing from this press release. They have to buy-in — literally.
TECH
Criteo is “live” on ChatGPT
A week ago, Criteo CEO Michael Komansinski announced that his ad tech company had partnered with OpenAI and will help facilitate ads in OpenAI’s ChatGPT. Read tipsheet. (March 3)
Last night, Chief Customer Officer Ed Dinichert announced on LinkedIn (March 10):
“You may have heard that we joined OpenAI’s advertising pilot in ChatGPT as their first advertising technology partner last week.
Update: We are live. Now.”
COMMERCE MEDIA
AppLovin mobile gaming consumer
Arguably, nobody should know more about mobile gaming consumer than mobile ad technology platform AppLovin. The company has feasted on the mobile in-app ad opportunity in recent years and built itself into a $160 billion advertising juggernaut.
On the blog for its AI-enabled ad platform Axon yesterday, AppLovin VP of Ecommerce, Cathy Sun, and Director of Product Partnerships, Shirley Deng, teased new research on the mobile gamer created in partnership with Kantar.
TL;DR: Gamers like to spend.
Given the company’s recently announced foray into eCommerce via its new self-serve Axon ad platform, which is awaiting broader release, the data showed continued eCommerce promise outside of the game genre.
A selection of purchase-related insights included:
- “70% of mobile gamers report that they make most household purchase decisions. That is significantly higher than the general population benchmark.”
- “They are financially confident and active online shoppers. 71% shop online at least weekly.
- “77% spend one hundred dollars or more per month online. Intent to purchase across high-value categories such as travel, electronics, furniture, and vehicles is strong.”
- “38% of mobile gamers report purchasing a product within three months of seeing an ad in a mobile game. Among those buyers, seventy-one percent acted the same day.”
Read: “The high-value mobile gaming consumer” (March 10) – Axon blog
Research: “Mobile Gaming: The New Mainstream Consumer Channel Report” (March 10) – AppLovin (bottom of page, PII required)
More: “AppLovin rebrands its ad platform as Axon, launches ads manager on referral-only basis” (October 2025) – ModernRetail
TECH
MadConnect aims to connect the agents
MadConnect CEO Bob Walczak spoke to AdExchanger about his company which identifies itself as an “intelligent connectivity layer” (ICL) for agents in service to marketers and agencies.
AdExchanger’s Allison Schiff reported:
“Under the hood, ICL is built to natively support Model Context Protocol (MCP), an open-source spec that standardizes how AI agents share context and interact across systems securely.
MCP defines how an agent talks to a data source or tool. But it doesn’t actually connect those systems to each other. That’s where ICL comes in, sort of like a switchboard behind the scenes.
If MCP is akin to a USB, think of ICL as the row of ports on a laptop.”
Read: “The Boring Infrastructure That Could Make Agentic AI Happen For Ad Tech” (March 10) – AdExchanger
Rio Longacre, Partner and Global AdTech Lead at Omnicom consultancy Credera, said of MadConnect’s ICL in a press release. “By establishing this neutral, platform-based layer, we can move past integration hurdles and focus on rapid innovation at scale, ensuring our clients’ campaigns are powered by the most agile and interoperable technology available.”
More: MadConnect Expands Its Intelligent Connectivity Layer to Power Agentic Marketing at Enterprise Scale (March 10) – press release
Related: PubMatic & Abovo Maxlead Deliver PubMatic’s First Agentic AI Advertising Campaigns in Europe (March 10) – PubMatic
LLMs & CHATBOTS
‘Ads in chatbots’ reviewed — WSJ
The latest edition of The Wall Street Journal’s AI & Business newsletter provided a skeptical overview of the emerging world of conversational advertising, aka “ads in a chatbot.” The short article’s premise is that ads will drive users away.
Nevertheless, new SimilarWeb stats show momentum for chatbot traffic and the promise it represents in ads and marketing:
“Even the more niche AI chat tools have sizable—and growing—audiences. Perplexity has just under 8 million daily active users now, which is nearly double what the “AI answer engine” pulled a year ago, according to Similarweb’s data.
Claude’s user growth, too, has been stratospheric despite Anthropic’s focus on enterprise and coding use cases. Claude ended February with about 9.4 million daily active users versus less than 2 million a year ago. And that doesn’t include a surge of 3.5 million users added just last week, according to Similarweb’s data.”
SimilarWeb data informed The WSJ that ChatGPT now has 250 million daily active users.
Read: “Why Ads in Chatbots May Not Click” (March 10) – The Wall Street Journal
From tipsheet: The story doesn’t really parse how Google’s AI Mode and AI Overviews fit into the chatbot ads equation. It’s still early for many looking to understand the evolution in AI, advertising and marketing — lots of education to come.
According to stats aggregated by analytics firm DemandSage in February, Google Search is estimated at ~5 billion users daily. AI Mode and AI Overviews plug into this estimated user base — let alone the fact that the Gemini LLM helps power search engine results.
Search is becoming a chatbot. Chatbots are becoming search. There will be no final “checkered flag” finish.
CONNECTED TV
Data point: Yahoo DSP’s pitch for CTV
“Reach over 190M+ monthly active Netflix viewers and tap into 696M+ user profiles to connect with audiences based on what they actually watch, search, read, and love.
Netflix gets the attention. Yahoo DSP helps you connect.”
That’s not a bad LinkedIn pitch for connected TV buyers still mired in reach and frequency requirements of the linear TV age.
Beyond that, Yahoo DSP’s AI-enabled tools would appear to be well-positioned to sift through the enormous scale.
Yahoo announced the ad tech unit’s Netflix partnership last June.
More: Yahoo DSP GM Adam Roodman’s tipsheet interview last month.
From tipsheet: Other AI-enabled DSPs funneling liquidity to Netflix inventory include…
- Amazon DSP which announced a similar deal to Yahoo DSP’s last September. Last week, Netflix added more addressability to its ad offering by integrating Amazon Audiences beginning in Q2 (read).
- In 2024, The Trade Desk (read) and Google DV360 (read related addressability announcement from 2025) became Netflix buyers with Magnite as the SSP intermediary in May of that year.
- Microsoft has been a buyer of Netflix inventory since 2022.
SELL-SIDE
Opinion: Publishers and agentic advertising
“We can discuss agentic advertising at a very high level: protocols, frameworks, architectures. But the moment someone asks ‘okay, what actually happens next?’, the conversation gets much more interesting.
So I wrote about it…”
Read: “Why publishers aren’t completely sold on agentic advertising” (March 10) – The Current
AGENCIES
Agencies testing facial recognition, AI
Media agencies such as indie Horizon and dentsu-owned Carat are beginning to test a consumer intelligence product which takes cues from facial recognition technology and AI, according to MediaPost’s Steve McClellan.
Looking under the hood of the tech partnership, he reported:
- “Vurvey Labs is an applied AI company that builds ‘people models’ and AI personas using video surveys to help brands gain authentic consumer insights. It accelerates product development, marketing, and co-creation by analyzing human feedback to create 93-95% accurate, AI-powered simulations of customer behavior.”
- ”Sightly is a marketing intelligence and technology company that uses its AI-powered Brand Mentality platform to help brands align marketing with real-time cultural moments. It enables brands to identify trends, mitigate risks, and activate across video and social media.“
Read more in MediaPost. (March 9)
PEOPLE MOVES
Now hiring
- Ellipse announced the appointment of former Auditude and Sharethrough sales leader Mike Gaffney as President (March 5) – press release
- Former TikTok Head of Growth Garland Hill joins MNTN as Chief Revenue Officer (March 10) – LinkedIn
MORE
- Google Ads adds AI voice-over to Performance Max video ads (March 9) – Search Engine Land
- About location fees for ads on Meta platforms (March 10) – Meta
- Perion’s AI Agent Outmax Now Available for TikTok, Delivering Up to 25% Greater Media Performance in Early Results (March 10) – Perion
- How Programmatic – And AI – Took Over The 2026 Winter Olympics, according to NBCU (March 10) – AdExchanger
- Smartly Expands Global Footprint with New Product, Engineering and Commercial Hub in Mexico City – Smartly