Roku announced a redesigned home screen aimed at deepening engagement across its more than 100 million users.
Variety’s Todd Spangler reported:
-
“On Wednesday, Roku took the wraps off a new home screen that it says is more dynamic, more personalized — and smarter. A key design point: It’s meant to help viewers more quickly find their next favorite show to binge-watch. The new Roku home screen begins rolling out Wednesday (May 27) across all Roku TVs and streaming devices in the U.S. The company says it will expand to additional countries in the coming months.”
Read: “Roku Unveils Biggest Home Screen Overhaul in More Than a Decade” (May 27) – Variety
The move mirrors broader shifts across the platform ecosystem toward AI-driven optimization and discovery:
- Netflix has been leaning harder into recommendation and conversational discovery. (1, 2, 3)
- Amazon is mixing shopping, video and ads. (1, 2)
- Meta continually optimizes feeds and advertising systems through AI-driven ranking. (1, 2, 3)
In Roku’s case, the transformation includes the TV home screen becoming less of a static menu and more of an optimization surface. It is a critical control point for the company.
The advertising implications include more engagement, recommendations, greater control over navigation, sponsored placement opportunities and more first-party viewing signals to feed the AI machine.
Read: Roku Unveils New TV Home Screen for the Next Era of Streaming (May 27) – Roku
From tipsheet on Roku: In an AI-driven media environment, the interface itself becomes a control point — where decisions get made.
Related: Why Critics Say Email-Based IDs Don’t Work For CTV (May 27) – AdExchanger
From tipsheet on IDs: CTV still lacks perfect signal resolution. Yet platforms are racing toward optimized interfaces anyway.
LLMs & CHATBOTS
Developments
- Economic Futures in the Age of AI (May 27) – OpenAI
- Google AI Threat Defense (May 27) – Google Cloud blog
- Meta launches Instagram, Facebook, and WhatsApp subscriptions, with more to come, including AI plans (May 27) – TechCrunch
CONNECTED TV
CTV startup Olyzon announces funding
Connected TV startup Olyzon, which positions itself as a “decisioning layer for CTV advertising,” announced a new round of funding.
Sir Martin Sorrell, Co-Founder of S4S Ventures and lead investor in Olyzon’s round, said on LinkedIn:
“Delighted to announce that S4S Ventures has led Olyzon’s $10M Series A. There is a great deal of noise in adtech right now. Acquisitions. Consolidation. Everyone repositioning for an agentic future.
Through all of it, one thing is clear to me: the infrastructure layer for CTV media buying is yet to be built. Olyzon is building it.
The agentic AI revolution and the consolidation of CTV onto programmatic infrastructure are two of the most important forces reshaping our industry. Olyzon sits precisely at that intersection.”
Read more. (May 27)
S4S Ventures is the venture arm of S4 Capital, an ad tech-focused agency holding company.
More:
- Sir Martin Sorrell’s S4S Ventures leads Olyzon’s $10m Series A to build the Agentic platform for CTV (May 27) – press release
- Pitch Deck: How Olyzon Sold Sir Martin Sorrell’s S4S Ventures on Its $10M CTV AI Bet (May 27) – Adweek (subscription)
From tipsheet: “Decisioning layer” may become one of the defining phrases of the agentic ad economy.
MEASUREMENT
Viewbotting as market distortion
Chief Commercial Officer Matt Barash of creative ad platform Nova Studio argues that the emerging issue of ‘viewbotting’ may represent more than traditional invalid traffic or creator fraud.
He defines viewbotting as “the artificial inflation of views, engagement, or audience signals to make content appear more popular or influential than they actually are. The obvious reaction is to call it fraud, and in many cases, that’s entirely fair. But from a media buyer perspective, the more interesting question is where the real distortion happens.”
Read more on LinkedIn. (May 27)
More: “WTF is viewbotting?” – Digiday (subscription)
From tipsheet: The challenge may also create opportunities for ad tech vendors alongside emerging “AI slop” detection systems.
In AI-driven advertising systems, fake engagement may become less of a fraud problem and more of a polluted training-data problem.
Related: IAB Tech Lab Releases Guidance On AI System Bot Management Strategies For Public Comment Until June 25 (May 27) – IAB Tech Lab
LLMs & CHATBOTS
The TAM for chatbot ad networks
Nic Baird, who is co-founder of Koah Labs, made his case for why his company’s ad network model for AI chatbots will win — there are a lotta chatbots.
On LinkedIn, he listed 12 chatbot apps with “serious user bases” in the Apple App Store that you may have never heard of.
His list begins with AI chatbots such as DeepAI, Luzia and Character[dot]AI.
Read them all. (May 27)
From tipsheet: The long tail of the chatbot ecosystem could become a huge opportunity — in fact, every commercial media property could become a chatbot host.
FINANCIALS
More AI winners emerge? Snowflake earnings
- Markets: Snowflake shares were up 37.5% in after-hours trading.
- Stat: “Revenue of $1.39 billion in the first quarter, representing 33% year-over-year growth.”
- Results: Snowflake Reports Financial Results for the First Quarter of Fiscal 2027 (May 27) – Snowflake
- CEO Sridhar Ramaswamy:“With Cortex Code and Snowflake Intelligence, we are extending from the trusted foundation for enterprise data and context to become the control plane for the Agentic Enterprise.”
- Read: Snowflake rockets 35% on earnings beat and plans to spend $6 billion on Amazon cloud (May 27) – CNBC
From tipsheet: The AI infrastructure trade may have a new winner — and Snowflake wants to be the control plane for enterprise agents.
Will ad tech be next?
BRANDS
OpenAI bets on two CMOs
On LinkedIn yesterday, Garett Sloane, senior editor of technology and AI at Ad Age, wrote:
-
“Is this the most coveted role in marketing—OpenAI CMO for consumers—or is it the riskiest? We’ll see, but OpenAI is splitting duties with its plan for a consumer-facing CMO, after naming Colin Fleming from ServiceNow as the B2B CMO at OpenAI, Ad Age’s Asa Hiken writes. We have been following OpenAI’s marketing strategy closely since it first ran a Super Bowl commercial in 2025, which was somewhat of a brand anthem that set the tone for using ChatGPT as a springboard for creativity…”
Read more. (May 27)
More: “A maturing OpenAI splits marketing duties between two CMO roles” (May 27) – Ad Age (subscription)
From tipsheet: Consumer AI branding and enterprise AI adoption are rapidly becoming two very different jobs for OpenAI. At least, that’s what it’s thinking right now.
EVENTS
Meta says RecSys matters
Meta will be producing a half-day event on June 3 titled, “LLMs for Recommendation Systems Forum.”
Meta researcher Luke Simon teases his opening keynote and the event’s premise:
-
“The recsys industry is at a familiar inflection point. A decade ago, the shift from linear models to deep learning was met with skepticism — today, the same debate surrounds the move to LLM-native recommendation systems. History shows that companies slow to adapt fall behind. This keynote explores why the convergence is happening now and what it means for the future of recommendations at scale.”
Other Meta participants include Sandeep Pandey, VP of Engineering, Adam Song, VP, Recommendation Research and Santanu Kolay, VP of Engineering.
You can register your interest for the event here.
Engineers from Amazon, Google, Spotify and Netflix are also participating.
Related: The generative RecSys revolution (May 27) – Eric Seufert on Mobile Dev Memo (subscription)
From tipsheet: Meta is effectively arguing that recommendation systems, rather than foundation models alone, remain a core control point in AI. The broader message is that recommendation quality depends not just on models, but on feedback loops, engagement signals and behavioral data that shape personalization at scale.
EVENTS
Meta’s Lattice model rollout targets apps
In a detailed note to investors, New Street Research’s Dan Salmon provided his takeaways on Meta’s Performance Marketing Summit which took place earlier this month.
He wrote:
“Performance Summit Key Takeaway: upgraded Lattice model rolling out now
The most important new detail from META’s Performance Marketing Summit was the rollout of an upgraded Lattice model, which began rolling out in 1Q26. Matt Steiner, VP of Monetization Infrastructure, Ranking & AI Foundations, said the latest version of Lattice is driving a 13% increase in click-through rate (CTR) and a 16% improvement in conversion rate on app event optimization.
The last part is meaningful because the metrics are for app-based advertisers only, i.e., mobile gaming, food delivery, online dating and advertisers focused on app downloads and/or re-engagement using products such as Advantage+ app campaigns. Steiner later notes that the upgraded model will be rolled out to other “objectives” and we believe this means that the upgraded Lattice model has not yet been rolled out to Sales and Lead Generation campaign types.”
Related: Ad models Overview —“AI Innovations in Meta’s Ad Ranking Driving Advertiser Performance” (2025)- Meta
From tipsheet: Meta continues to deepen its closed-loop AI advertising infrastructure. The apparent sequencing of the upgraded Lattice rollout toward app advertisers suggests Meta may be prioritizing environments with the strongest feedback loops and clearest conversion signals before expanding the model across broader campaign objectives.
LLMs & CHATBOTS
OpenAI to self-serve advertisers: Set the pixel
OpenAI is requiring advertisers to implement its Pixel or Conversions API if they want early access to the wider rollout of ChatGPT’s self-serve ad platform next week.
A note sent to self-serve advertisers (like me) yesterday said: “In early June, we’ll begin offering conversion-optimized campaigns, with early access for customers who set up Pixel or Conversions API in advance. Any account that has conversions set up by Monday, June 1st, will be granted early access by June 5th.“
From tipsheet: That’s notable. Historically, new ad platforms were happy to take your money whether your measurement setup worked or not. OpenAI appears to be taking a different approach by effectively telling advertisers: set up proper conversion tracking first.
OpenAI is effectively saying this is optimization-first architecture.
MORE
- The Trade Desk’s Fall From $69 Billion Digital-Ad Darling (May 27) – The Wall Street Journal (subscription)
- How Amazon Web Services is helping retailers build their own AI-powered shopping experiences (May 27) – press release
- Cadent Launches AI-enabled Predictive Audience Builder, Giving Advertisers Direct Control on How Audiences are Actually Constructed (May 27) – press release
- AdRoll Launches MCP Server to Make Campaign Insights and Actions Available Inside AI Tools (May 27) – press release