Here’s what marketers want from Meta’s Superintelligence Labs
By Krystal Scanlon • August 19, 2025 •
Ivy Liu
Meta’s Superintelligence Group has started turning up in marketer pitches. So far, its generating curiosity — and confusion.
As Chris Matheson, media director at Markacy pointed out, marketers and investors will be looking for proof of return on investment. “With $64 to $72 billion in 2025 capex, brands expect tools that lower customer acquisition cost (CAC) or raise conversion rate (CVR), not just research headlines,” he said.
Officially launched last month (July 1), Superintelligence Labs consolidates Meta’s core AI teams — FAIR, foundational models and product — all under one roof. And while there are tons of ideas on a typical marketers’ Meta wishlist, here are the most common things they said they’d like to see from the company’s new Superintelligence Labs.
Better control and transparency
Meta’s Advantage+ (among others) has come under fire for being a set-it-and-forget-it AI-powered campaign tool, which ultimately requires marketers to trust what it does, not to ask questions and just accept the outcomes. The result is a lack of transparency, the kind that earned these platforms the nickname “AI black box”.
That’s the tension. If Meta wants advertisers to keep spending. It has to give them more control – and visibility into how the machine makes decisions on their behalf.
“That’s a really common ask from our clients, and us as well,” said Kevin Goodwin, svp, strategy and growth at New Engen. “They want to know why. They want to understand it. They want to be able to explain it to leadership. So transparency is a huge ask and demand, especially from some of our largest clients.”
These days, social platforms are under far more scrutiny by regulators and watchdogs when it comes to what data they collect from its users, and in what ways. Marketers will want assurances on the data collected, processed and stored.
“It’d be huge if Meta’s AI could infer customer demographics, purchase intent and interest segments in a privacy-safe way,” said Brian Wong, CEO and co-founder of Feel Goods. “Even if we don’t see the raw data, having AI surface patterns like “this creative is resonating most with Gen Z women who engage in wellness content” would massively improve our targeting efficiency.”
For the most part, Meta has focused on lower funnel advertising. And while advertisers see value in performance, they also want to build brands.
“Meta’s has always been a direct response channel, and really invests heavily into products that fuel success within that realm,” said Goodwin. “I haven’t seen AI or these other areas as attached to the full funnel or the upper funnel, and so I would love to see more prioritization of that.”
Predictive creative guidance
With AI flooding the ad market, the real differentiator for brands isn’t the tools – it’s the creative they put into them.
That’s where Meta’s Superintelligence could matter. The pitch: not just analyzing what worked in past campaigns, but predicting which creative is most likely to work going forward.
“This would give us a roadmap of creative angles, formats and even influencer styles before we spend big on testing,” said Wong. “Especially for a lean team, that would save an incredible amount of wasted spend and content cycles.”
Given the headlines claiming Meta will automate the full ad process by the end of next year, this doesn’t seem like a ridiculous idea at all.
Demonstrate a real competitive edge
While Meta is a leader in the platform space, it’s yet to demonstrate the same trailblazing levels when it comes to AI. Which probably explains why it’s trying to shortcut its way to the top by poaching a number of the top AI specialists from its main competitors. Think OpenAI’s chatGPT co-creator Shengjia Zhao, Apple’s former head of large language model (LLM) development, Ruoming Pang, along with Google DeepMind researchers Jack Rae and Pei Sun.
Why the hesitancy from the market? Because it wasn’t so long ago that the industry saw then-Facebook rebrand the entire business to Meta when it went all-in on the metaverse. Sure, it’s probably spot on in thinking AI is the way to go, but the industry wants to see this isn’t just another fad.
“Advertisers and investors are watching closely as Meta pours billions [of dollars] into AI infrastructure and talent, with little short-term clarity on monetization,” said Markacy ‘s Matheson. “Meta’s Superintelligence Lab must [demonstrate it can] outpace Google, TikTok and OpenAI on usable capabilities, smarter optimization, faster learnings and real-time attribution.”
Meta did not respond to Digiday’s request for comment.
