From Amanda
Meta keeps adding more AI to advertising — here's what I'd actually pay attention to
Every week there's another Meta AI announcement. Between reports that Meta may eventually sell AI compute, this week's launch of Muse inside Advantage+, and plans to manufacture its own AI chips, it's easy to wonder if any of this matters. You don't need to understand custom silicon or gigawatts of compute. What matters is that Meta clearly believes AI is becoming the foundation of its lucrative ad business.
Muse is the latest example. Meta built it directly into Advantage+, where it can generate and refine creative variations based on an advertiser's existing assets and creative brief. Advertising on Meta is already evolving in this direction — more than 8 million advertisers are using at least one of Meta's generative AI creative tools.
That's why so many of the updates over the last year have centered around creative generation, audience expansion, placements, bidding, and optimization. The same massive investment happening behind the scenes powers all of them.
It also means marketers have fewer levers to pull inside the platform than they used to. If you've ever launched a campaign for a brand-new app with a limited budget, you've probably felt this firsthand. Meta works best when it has enough data to learn from, and that's often the hardest part for smaller advertisers.
It's also changing how marketers spend their time. Instead of trying to produce one perfect ad, you'll likely spend more time developing creative concepts, writing stronger creative briefs, producing high-quality assets, and giving Meta's models more to work with. The platform is increasingly responsible for generating, testing, and optimizing creative variations on your behalf.
Regardless of your company size or budget, it’s a good idea to get comfortable with these models. Meta is putting AI at the center of campaign creation and optimization, and Google has been moving in the same direction. Advertising effectively on these platforms increasingly means learning how to work with AI rather than around it.
But all of this makes it even more important to step back and think about where Meta fits into your overall marketing mix. AI can optimize campaigns inside Meta, but it can't tell you whether another dollar belongs on Meta, Apple Search Ads, TikTok, connected TV (CTV), or somewhere else.
And if everyone's using Advantage+ and Muse, the quality of the thinking and creative inputs behind them become even more important. Mick Rigby of Yodel Mobile recently called this "the great flattening." When the whole industry is optimizing with the same systems, it's worth remembering that the algorithm alone can't decide what makes your brand worth paying attention to. That’s still the marketer’s job.
Advertising
The CTV acquisition spree is really about measurement
If you've been following CTV news lately, it feels like there's another acquisition every time you blink. Walmart bought Vibe.co. Fox is buying Roku for $22 billion. Viant bought TVision. Pinterest bought tvScientific. Meta is reportedly in talks with supply-side platforms (SSPs) to get into CTV too.
None of these companies are simply buying access to television inventory. They’re buying into capabilities that make television perform more like digital advertising. For years, CTV was about reach and impressions — a branding channel where you measured awareness and hoped the results showed up somewhere else.
Now, brands want to know if a CTV ad drove an install, a purchase, or incremental revenue. Walmart's acquisition of Vibe, following its purchase of Vizio, gives it the ability to connect commerce data, media buying, measurement, and optimization into a single feedback loop. Instead of simply reporting whether a campaign worked after the fact, it can increasingly optimize campaigns against real business outcomes while they're still running. Viant's purchase of TVision was an even more direct bet on attention measurement. Pinterest's acquisition of tvScientific brought intent signals into CTV campaigns. Even Meta's reported interest in CTV indicates attempts to bring its performance advertising engine to the biggest screen in the house.
A recent survey from Jamloop found that 50% of marketers increased their CTV budgets this year, yet only 33% fully trust platform-reported performance claims. More than 70% said they'd invest even more if measurement and attribution improved.
CTV is starting to win in the budget conversation, but measurement is still a gap that is going to be the big difference-maker here. It’s also why I think all of these acquisitions are popping up. They signal a broader shift across the industry: Companies are investing in the data and measurement capabilities needed to connect CTV to real business outcomes.
So when we take a step back, nearly every major player has reached the same conclusion: CTV is becoming a performance channel, and measurement is critical to its next phase of growth. For mobile marketers, that's especially important. Many of the outcomes brands actually care about happen on mobile, even if the customer journey starts on the TV screen. If you can't connect those two touchpoints, you're still left making an educated guess about CTV's impact.
Spotlight
User-generated content (UGC) is having a moment (and it's not an accident)
You can vibe code an app in a weekend now. What you can't shortcut is getting anyone to care that it exists. That's a theme that keeps coming up in conversations I've had recently: As building gets easier, distribution gets harder — and UGC is emerging as one of the more interesting ways to solve it.
David Barnard, growth advocate at RevenueCat, specifically called out being great at UGC or managing UGC as one of the distribution advantages he's seeing among successful app founders. He shared a story about vibe coding an app inspired by his own kids about building character, soon realizing the perfect person to do UCG for it was his wife, who also happens to be a licensed therapist.
He'd started noticing a pattern in what actually performs: creators leading with real credentials — "I'm a licensed therapist, here are the lessons I wish I knew" — before delivering advice. His takeaway wasn't to chase the biggest following. As he put it, "You're buying a creator, not an audience," and the creator worth buying is "an expert not just at the field, but an expert at creating content."
Nic Weber, co-founder at Noise, shared a framework that helps explain why the right creator and message can work so well: relatability, need, and novelty.
Relatability is the "I see myself in this" factor — and he thinks it's specifically why expertise-driven content is outperforming polished influencer marketing right now. His read: Audiences have grown wary of aspirational, overly polished content, and trust people who look and sound more like themselves instead. Need is the problem someone already wants to solve. Novelty is what makes the solution interesting enough to pay attention to.
AI is making it dramatically easier to create more products and more content, but it hasn't made human attention any less finite. If anything, attention and trust are becoming more valuable as more things compete for them.
More and more consumer brands are making UGC a core part of their strategy to lower acquisition costs and improve conversion across the customer journey. And data from this year suggests it works: UGC drove 6.7x higher conversion rates than non-user-generated content.
I think UGC is having a moment because when it's done well, it combines real expertise, relatability, and creator skill in a way that can actually earn attention.
Adam's Take
The arc of change for AI-enabled marketing
We live in interesting times, and some of the most interesting stories that lead the news today concern how AI is changing product discovery for consumers. We read — and experience — how consumers now operate within the confines of an LLM but we have yet to see the downstream impact of AI on how companies operate in this new era.
Chiefmartech has a wonderful yearly primer on the state of marketing, and this year it did a fascinating projection on how AI-enabled marketing will change company operations.
It makes an interesting observation that when the customer is living within the walls of an LLM, providing proper buyer context isn’t aspirational; it’s the entire marketing journey.
This change will manifest itself as starting with a messy, but necessary, “AI everywhere, integrated nowhere” Frankenstein approach, but will mature into an end state where customer agents can operate independently and marketing’s job becomes providing customers “context-as-a-service.” The result will be a truly dynamic customer journey — but one completely controlled by the customer(‘s agent).
AppsFlyer’s investment is a nod to the MMP’s potential future
AppsFlyer’s recent billion-dollar cash infusion wasn’t really a surprise. After an off-again, on-again IPO and private equity dance, it was clear they were chasing liquidity. And really, how can you blame them? The company was founded 15 years ago, and longtime employees are undoubtedly itching to free themselves from the shackles of early-equity options.
What was a surprise were the investors: Google, Meta, Moloco, and Unity. Customers rely on AppsFlyer to measure and adjudicate measurement signal coming from these exact platforms, so a conflict of interest presents a looming worry. This tension explains why AppsFlyer has been so adamant about pledging ongoing neutrality to its customers.
Conflict of interest aside, this investment reveals something more interesting: The ad platforms are betting that MMPs will become essential infrastructure across all advertising channels, not just mobile. Will the MMP remain relegated to the mobile app ecosystem? Or is there real potential to serve a larger domain?
Compared to mobile, web measurement is relatively easy. The straightforward nature of URL and UTM logging allows Google Analytics to offer a (largely) free product that monopolizes measurement in the web analytics space. MMPs exist because of the sheer complexity in measuring and deduplicating advertising signal in the mobile app space, which means that today, most companies with web and app presences use separate tracking technologies. But the macro landscape of marketing isn’t getting more straightforward: Connected TV, social, search, and email create multiple customer touchpoints that explode the number of measurement signals advertisers and marketers need to collect and evaluate. And AI advertising will just make this worse.
Advertising in an LLM is shaping up to look more like mobile than web. The space is dominated by a few isolated walled gardens that own and manage their own advertising platforms, tightly control user experience, and deliver measurement signals back to advertisers. This is pretty darn exciting if you’re an MMP. If AI advertising is a new explosive growth vector (it is), and the platforms are sending out advertising measurement signal (they are), then marketers need tools to parse through and measure advertising impact. This is exactly what an MMP does.
Investment from the large ad platforms into an MMP is a strong indicator that the ad platforms see value in independent measurement and want to expand the role of MMPs. Now we’ll just have to see if AI advertising scales (it will).