Future of Marketing: Brands want creators who can win over humans and machines at once

Three agencies working this space right now can each point to the moment clients stopped treating it as theoretical. With Reign Maker, AI visibility has been the first or second thing raised with every client for six months straight running. Over at Go Fish Digital, it shows up as a steady stream of clients asking how their creator content is landing across the most cited platforms. Ogilvy is further along with some accounts: a handful of its more advanced clients are already setting explicit KPIs around discoverability, credibility and culture, rather than waiting to see how the conversation plays out.  

Some of the results so far challenge a lot of what marketers assume about creators. 

“It’s not necessarily built on engagement rates or follower counts,” said Lauren Lyster, vp and head of social media at Go Fish Digital. “What we’re seeing probably get cited is a lot more long-form information. We’re seeing that outperform Shorts. They don’t often actually get cited very often in AI data.”

Strip away the specifics and it all traces back to one thing: marketers spent years learning to talk to people. Now they have to talk to the machines recommending them too. And creators, sitting on exactly the kind of firsthand, credible content those machines are built to cite, have become one of the few reliable ways to do both at once.  

Tinuiti is already watching that shift happen in real time. Six to eight months ago, AI visibility was maybe a tenth of the factors the agency weighs when building out a creator strategy. Now it’s closer to a quarter, and the conversation itself, clients asking how to do it, touches something like 60% of the agency’s roster. All of that stays contained to strategy, though. Rate cards haven’t moved, and creators are folding citations into media kits and pitches as a way to stand out rather than a reason to charge more. Execs expect that to change, just not for another six to 12 months, once brands get a firmer read on what a citation is actually worth. 

Until then, Tinuiti is using the data to sharpen how it briefs and deploys creators. It’s one of the first questions the agency asks now: what’s the influencer strategy because the answer feeds straight into the keywords and phrasing that go into a brief. It’s also nudging the creator mix itself, toward voices already showing up in the citations that matter for a given category, rather than picking purely on follower count. The data, in other words, is setting the direction not the price.

“I wouldn’t even say acute anymore,” said Crystal Duncan, evp of brand engagement at Tinuiti of how often the topic comes up with clients now. “It’s front and center right now.”

Execs at Dept can attest to that. Before a campaign starts, they check whether a creator is already in AI-generated answers for their category then weigh that alongside the same factors that have always made someone worth hiring — category authority, credible content and audience trust. What actually changes shows up later in the brief itself. On one recent engagement, that meant handing creators a reference list of exact product names and verifiable claims to script from, structuring chapter markers around the real questions people ask rather than plain timestamps, and insisting on a clean transcript instead of letting the model parse error-riddled auto-captions. 

“We still want [creator] content to be natural, in their own voice, saying what they want to say,” said Angela Seits, Dept’s vp strategy. “But making sure the structure behind the content includes some of those important influences on LLM discoverability.”

The Now Agency isn’t coring creators on this yet but it is looking for something more specific than reach: whether a creator has built up real domain depth that a model would treat as authoritative. Its benchmark is consistency over time in one lane. Or as Gabe Feldman, co-founder and managing partner of the agency put it. “We wouldn’t necessarily look at somebody that, for five years, has been producing beauty and skincare content that is now talking about coding and engineering. We’d look at somebody that has deep-rooted domain expertise in a very particular category and has done that consistently and constantly for a very long time.

None of this should be much of a surprise. Referral traffic from search is drying up, zero-click answers are becoming the default, and the models filling that gap lean on the same reference points that creators have spent years accumulating without ever trying to. Most of that groundswell sits on YouTube right now. Every major LLM treats it as a key citation source nowadays, and that’s quietly rewired what marketers and creators are optimizing for. 

But the shift runs deeper than platform preference. Creators used to answer to reach, engagement and how much of a campaign they could carry. Today, a second system runs alongside that, one that treats a creator’s opinions and demonstrations as evidence a model can cite — evidence that keeps working long after the campaign that paid for it has ended. Under those rules, a creator with 100,000 followers and real depth on a subject can outweigh one with five million followers. It has the sort of authority a bigger account without that depth, simply can’t fake.

“We as marketers have to look at targeting humans and machines,” said Justine Stauffer, said senior director of creative effectiveness at Molson Coors. “We also know the role of creators and the role of organic social really helps the machines understand and ingest what our brands are all about and all of the right occasions that people want to be able to enjoy with our brands and choose us to hang out with.”

Jellyfish is seeing that authority question lay out at scale across its own client base. More than 90% of the marketers who employ Jellyfish now list AI’s impact on discovery and sales as a top three concern, up from a year ago when, in vp of strategy John Dawson’s words, “many, many clients had no real drive around GEO or AEO” at all. Narrow that to creator programs specifically and its roughly half of advertisers with an existing program who are already looking at this. 

The agency’s answer is “Share of Model”, a tool built to measure that shift and then act on it. It runs large batches of prompts against the major LLMs and tracks which sources get cited across tens of thousands of results. From there, AI analyzes the cited videos themselves, pulling out what’s actually driving the pickup so clients get a read on why something got cited, not just that it did. That feeds straight into the creator’s brief, which formats a category that responds to how long a video should run and what a model reads as signal versus what a human viewer notices. The output, per Dawson, is a “specific priority plan” that spans areas like who to work with, what to ask them to make and how to frame it around the questions people are actually putting to the models.  

“How does this shift the portfolio mix of creators that I use and the sorts of remuneration models as well,” he said of those questions. “If you’re looking for citations and not views, creators as well have the ability to think about different ways of running things.”

Well, maybe not too different — for some at least. Plenty of the activity in this space still looks more like positioning that practice. Look at the agencies bolting AI language onto existing creator work to be seen ahead of the curve without a settled answer to what the tool is actually for beyond appearing cited. The bigger risk, then, could be over-engineering a problem that mostly solves itself. Content that’s helpful, entertaining and ownable was already good creator work before anyone thought to measure it against a model. 

“It’s not like I can stick a million quid into Google and see my SEO rankings change overnight,” said Charlie Coney, creative and strategy officer for Ogilvy’s U.K and EMEA regions. “LLMs take longer to train, and the way those algorithms work together isn’t hugely public so there’s a bit of guesswork in here.” 

Numbers to know

$1 billion: The annualized run rate that OpenAI’s advertising revenue has already achieved.

14: Number of days until Judge Brinkema’s reasoning for her verdict on Google becomes unsealed.

23%: Percentage of U.S. shoppers that rank AI assistants as trusted sources for shopping, behind family / friends, professionals, retailer reviews and influencers.

$18 billion: the figure Meta settled its U.S. lawsuits over children’s social media addiction for.

What we’re reading

TikTok Backs Out of House Meeting to Avoid Kids Safety Questions

TikTok pulled out of a congressional meeting on its U.S. operations, citing its ongoing child safety litigation, according to Bloomberg. But the committee chairman says the meeting was about national security, data and Chinese influence — and argues TikTok’s withdrawal raises fresh questions about its transparency and ByteDance’s continued influence.

Meta debuts Muse Spark 1.3 as personal agent work continues

Meta has released Muse Spark 1.3, an AI model upgrade focused on coding and agentic tasks, as it tries to keep pace with OpenAI, Google and Anthropic. Meta says the improvements are laying the groundwork for personal AI agents that can work on users’ behalf 24/7, while the company is increasing safety investment without pausing development, per Axios.

X continues to lose EU users

X is continuing to lose users in the EU, with its latest Digital Services Act disclosure showing the decline has persisted since 2023, including a roughly 15% drop in the second half of 2025, per Social Media Today. The decline comes alongside a 70% fall in ad revenue versus the first half of 2022, though X’s value to xAI as a source of user data could make the shrinking audience less damaging to Musk’s broader business.

Social media companies are trying harder to know which users are kids

Meta and other tech companies are expanding AI-powered age checks to keep children off adult platforms, but privacy advocates warn the technology could compromise users’ privacy and anonymity, per AP News.

What we’ve covered

OpenAI’s ChatGPT ads business hits $1 billion run rate as Europe gets self-serve access

OpenAI’s ads business hit a $1 billion annualised run rate in under 200 days, with self-serve Ads Manager beta now open across 31 European markets. But the run rate flatters things — only about $330 million is actually booked this year, so the $2.5 billion targets needs a steep Q4 acceleration.

WTF is the Amazon ad surcharge lawsuit?

The FTC and 22 states allege Amazon sold second-price auctions but from 2019 added a hidden markup, so advertisers paid near their full bid — allegedly 80% of the time by 2024. Over a million sellers affected; Amazon denies anyone was overcharged. Notable because it’s the first regulatory challenge to auction mechanics inside a closed retail media network.

Why Google’s ad business escaped a breakup

Judge Brinkema spared Google’s ad business a breakup, ordering behavioral remedies instead — curbs on self-preferencing, publisher data-sharing, fair treatment of rivals. Her reasoning stays sealed for 14 days. The outcome is predictable given her doubts about finding a buyer and a market moving too fast, but behavioral fixes are equally hard to enforce, and Google will contest every clause.

How every major platform pays creators

A reference guide to monetization thresholds, payout minimums and eligibility rules across Instagram, TikTok, YouTube, Facebook, X, Snap and Twitch. The through line: the goalposts keep moving — YouTube doubles its requirements in February 2027, X replaces revenue sharing on Sept. 7, and Twitch now gives everyone fan-funding tools but holds payouts until Affiliate.