January 2nd-January 8th // Estimated Reading Time: 9 minutes
In This Edition!
Meta drops $2B on Manus
Walmart’s Sparky (also) barks ads!
Press Release: Agency Holding Company Launches All-In-One Marketing AI Platform So No One Has To Ever Do Anything Again
The next AI Accelerator starts NEXT WEEK! We still have a few seats left.
ENROLL NOW!
Email [email protected] for team discounts.
#KnowledgeIsPower #HumansAheadOfAI
Top Headlines🔥
Meta drops $2B on Manus, a Singapore-based AI agent startup. Now, China is reviewing the deal.
Source: NBCU, FT
December 29th, 2025-January 6th, 2026
Summary: At the very end of 2025, Meta acquired Manus, a startup that builds AI agents that can handle complex tasks like market research, coding, and data analysis without human intervention. Within eight months of launching, Manus reached $100M in annualized revenue (fastest company to $100M in ARR ever) and now serves millions of users globally. Meta will integrate Manus's technology into its products, including the Meta AI assistant, and attempt to scale the service by cross selling it to its users and customers. However, because Manus was launched in China before relocating to Singapore, China is reviewing the deal🔒 to determine if Manus needed an export license under Chinese law for the move and sale.
Opinion: Manus is cool! But also, Meta's AI strategy seems a bit all over the place. It’s building AI models Llama and Avocado in-house, investing $14B in Scale AI, acquiring wearables startup Limitless, and now dropping $2B on Manus. With investors getting nervous about how much Meta is spending on AI, at least Manus can contribute some revenue to the bottom line. Still, Meta’s moves give the impression that it is trying to hedge every bet simultaneously. While its big tech rivals, like Google, Microsoft, and Amazon, have very focused AI strategies that are paying immediate dividends.
Walmart brings ads to Sparky AI shopping agent, unleashes Marty AI for advertisers🔒
Source: Adweek
January 6th, 2026
Summary: Walmart is testing ads within Sparky, its AI shopping assistant that lives in the Walmart mobile app. When shoppers ask Sparky for product recommendations, they may see sponsored prompts alongside organic results. Walmart says 81% of customers have used Sparky to check product availability and details before buying. The move puts Walmart in direct competition with Amazon's Rufus, which started running ads in 2024. Walmart is also expanding Marty, its AI agent that helps advertisers manage campaigns. Marty will be available to all Walmart advertisers buying search ads in the first half of 2026 and will expand to other ad formats throughout the year.
Opinion: With retailers turning AI shopping assistants into ad platforms, they now have AI agents on both sides: Sparky decides which products to recommend to shoppers, while Marty decides how to optimize your ad campaigns. Inevitably, Sparky and Marty will talk to each other. They'll be in cahoots with each other. This may give us a glimpse into the future: Retailer-controlled AI agents negotiate with each other over which products get surfaced, organically and in ads, which may give brands less control over how their products show up and which levers to pull in order to optimize.
Holdco AI arms race accelerates as Havas, WPP, Omnicom unveil arsenals
Source: Various
January 5th-7th, 2026
Summary: Three large agency holding companies unveiled major AI platforms within 48 hours at CES, signaling a concerted push to control how AI reshapes agency work. Havas launched AVA🔒, a platform designed to turn its 25,000 employees into "product engineers" who can build AI applications without coding, while also exploring a shift from hourly billing to value-based contracts. WPP debuted Agent Hub within its WPP Open platform, offering marketers curated "super agents" that tap into three decades of proprietary data across branding, creative, and behavioral science. And Omnicom updated Omni, its marketing intelligence platform, to integrate all of IPG’s assets and augment its AI capabilities.
Opinion: Wow! The holdcos are really crushing it with AI launches and press releases. But … what’s actually going on behind the scenes? If you work at one of these companies, are you really USING any of these tools? If you’re a client of these agencies, are you seeing the benefits and cost / speed efficiencies? How much of this tech is actually tangible at this point? Hit reply and let us know. We can keep you anonymous! (unless you don’t want us to 🤣)
New Products & Features 🚀
What It Does: Max Campaigns is an ad buying tool that uses AI to optimize targeting, creative selection, and budget allocation. It takes an "open box" approach—transparently showing advertisers granular audience insights such as how individual creative elements perform. Some early testers saw 17% lower cost per acquisition and 27% more conversions on average.
Quick Take: It's Reddit’s turn to launch a PerformancePlusAdvantageMaxCampaignPlus+ AI product to rival the other walled gardens! The "open box" positioning addresses advertiser frustrations with other black-box AI tools. But, at the end of the day, it'll come down to whether Reddit's conversation data can help brands understand audiences in ways display and other social platforms can't, and ultimately drive better outcomes.
What It Does: Independent DSP Viant's new AI-powered tool automatically runs and optimizes ad campaigns without human involvement, using Viant's data on households, content context, and ad quality to make real-time buying decisions. Underpinning the solution is “AI Lattice Brain”, a decisioning architecture Viant has developed for autonomous campaigns. Advertisers can see exactly where their ads are running and how they're performing.
Quick Take: Like Reddit, Viant has also read the room, betting that advertisers want Google and Meta-level automation without the mystery. But the same questions applies: whether transparency alone can compete with the massive scale and data advantages of the big platform. Outcomes matter most. Early performance looks promising: Viant's AI delivered 2.3x better cost per action than a human🔒.
What It Does: Magnite embedded an AI agent into SpringServe, its CTV ad server, to test agent-to-agent communications. The idea: A buyer's AI agent describes what kind of campaign it wants, and SpringServe’s AI agent matches it with ad inventory and audiences.
Quick Take: We always talk about agents BUYING ads. But agentic SELLING will be important too. It will give early adopter publishers and supply side tech an advantage. Smart of Magnite to jump on it.
What It Does: Optable's AI agent, built on Claude, reads advertiser RFPs, generates audience recommendations, and activates campaigns—reducing the RFP response process from weeks to hours.
Quick Take: In the near term, speed becomes a competitive differentiator. Publishers who can turn around RFPs in hours instead of weeks will win budgets before slower competitors. That said, eventually, speed will become table stakes.
What It Does: LiveRamp's Data Marketplace now lets marketers license three types of AI assets: permissioned datasets to train and tune models; third-party AI models that run on your first-party data without exposing it; and AI-powered applications and agents for audience building, measurement, and media optimization (coming soon).
Quick Take: Is LiveRamp becoming an app store for AI? It's a cool concept. A centralized marketplace where companies can discover and license the models and agents they need for various marketing use cases rather than building everything in-house.
AI Use Case of The Week💡
The Setup: Bank of America wanted to scale personalized financial guidance and value across 50M users.
The AI Solution: Bank of America trained Erica with a library of 700+ responses to client questions. The team made 75,000+ updates since 2018, continuously training the system to handle new customer needs while maintaining accuracy. Erica's capabilities have been scaled across the bank's business, including consumer banking, wealth management (Merrill), corporate banking (CashPro), and internal IT support.
The Results:
• 3B total interactions since 2018, now averaging 58M monthly
• 98% of users are able to find the information they need, slashing call center volume
• 50%-60% of interactions are now proactive, where Erica suggests actions and users engage
• 90% employee adoption internally, reducing IT help desk calls by 50%
Why It Matters: Generative AI is all the rage, but a deterministic approach—using AI to understand customer questions, then pulling from company-approved responses—can be useful as well, especially in tight regulatory environments. Every response can be audited and vetted before customers see it. And there can be many, hundreds of thousands of responses at the ready.
Your Action: Eat your own dog food. Before building customer-facing AI, deploy it internally first with your own employees to create real-world feedback loops and internal champions. Also, consider designing your AI architecture for reusability from day one. Ask "which other business units could use this core capability with minor modifications?" rather than building AI point solutions for each use case.
Other Notable Headlines📌
OpenAI’s ads push starts taking shape🔒- OpenAI employees have been exploring the mechanics of how to inject advertising into ChatGPT without alienating users who expect unbiased responses.
AI agents are taking over NBCU's linear TV buys - NBCU, RPA, and FreeWheel are executing the agent-to-agent ad buys for live sports across linear and streaming, with AI agents handling everything from RFPs to creative approvals in seconds for Q1 2026 NFL playoff campaigns.
Yahoo brings AI agents to its DSP - Yahoo DSP's new hybrid approach lets advertisers use its AI agents for media buying, troubleshooting, and other tasks, bring their own AI agents, or use a combination of both.
IAB Tech Lab Releases Agentic Roadmap to Scale AI Buying - IAB will integrate modern agentic frameworks into existing protocols like OpenRTB and AdCOM rather than rebuilding from scratch.
Shoppers turned to AI this holiday season—but mostly for customer service🔒- While AI influenced 20% of global retail sales, only 2% of shoppers actually use AI assistants to start product searches. But 78% interacted with AI-powered customer service, with usage jumping 126% during November and December compared to the prior two months.
Almond Breeze jabs at AI-generated slop with the Jonas Brothers - AI criticism is increasingly part some brands’ messaging as they try to convey authenticity and human connection.
Stagwell agency Code and Theory introduced at CES “The “Machine,”🔒- It's an agentic operating system that automates marketers' existing AI tools. PubMatic also debuted an agentic operating system to orchestrate multiple AI agents with launch partners WPP Media, Butler/Till, Wpromote, and MiQ. An agent to control all the agents!
That’s It For This Week 👋
The U of Digital AI Newsletter is intended for subscribers, but occasional forwarding is okay!
To subscribe visit Uof.Digital/Newsletters or contact us directly for group subscriptions.
And remember, U of Digital helps teams drive better outcomes through structured education on critical topics in marketing / advertising like AI, ad tech, mar tech, privacy / identity, CTV, commerce media, and more. Interested in learning more about how we can supercharge your team? Let’s talk!
Thanks for reading!
