In mid-June at Cannes Lions, agency leaders took the stage one after another to say the same thing: the AI hype cycle is over, craft is back. Backstage at the same festival, Luma AI was handing out a $1 million prize for the best ad made with AI. That same week, Nielsen's global report showed trust in advertising jumping eight points, from 39% in 2025 to 47%. And in the same week, Meta raised its 2026 capex guidance to $125-145 billion, the largest automation bet in the company's history, announced right as "craft won" became the headline.
All three things are true, all three happened in the same seven days, and all three contradict each other. A festival stage applauded craft. A balance sheet funded automation. A trust survey didn't mention either, the eight-point jump traced back to a 63% surge in verified disclosure labels. The real story of 2026 isn't "AI versus craft." It's three separate things happening at once, and none of them substitutes for the others.
Why doesn't craft read as a trust signal anymore?
Because generative tools raised the production floor so high that "looking well-made" no longer proves anything. A decade ago, high production value signaled that real money stood behind a brand. Today any small business reaches the same polish in two hours for close to nothing. McKinsey's newest consumer research backs this from a different angle: shoppers now decide who earns their attention faster and more skeptically, with less patience for generic messaging than a year ago. The brain still makes the trust call fast and on instinct, but the cue feeding that instinct has moved. Audiences stopped asking "how well was this made" and started asking "what didn't you tell me." Craft stopped being a visual signal the moment everyone could buy the same visual quality. Disclosure became the verifiable one, because a label is binary, it's there or it isn't, and leaves no room for a debate about taste.
What did Nielsen's eight points actually measure?
Not a creative breakthrough, an administrative one. The share of brands clearly labeling sponsored content, AI-generated material, and sourcing information rose 63% in a single year. Millennials and Gen X trust advertising the most; Gen Z and the 65-plus cohort trust it the least, and that tracks with a separate finding from Gallup: fewer than one in six Gen Z respondents trust big tech companies at all. So labeling doesn't move everyone equally, but it moves the average, and it moved it more than any creative trend measured this year. The uncomfortable part is how cheap that lever is. The fastest way to buy trust in 2026 wasn't making a better ad. It was putting the right disclosure box on the ad you already had.
So why does the budget still go to automation, not craft?
Because the numbers close the argument before it opens. Meta's Advantage+ system returns $4.52 for every dollar spent and on its own generates roughly $60 billion in annualized revenue, in a quarter where price per ad rose 12% and impressions rose 14% at the same time. No craft deck walks into a finance meeting and beats that table. The $125-145 billion in capex isn't chasing the thing that got the standing ovation on stage, it's chasing the thing nobody mentioned from the podium, because CFOs don't sit in the Cannes audience. They sit in the quarterly call, and the only metric that room rewards is return.
- Fund disclosure labeling as an actual line item. Treat sponsored, AI-generated, and sourcing disclosures as a measured investment, not a compliance afterthought.
- Spend against your own KPIs, not the Cannes headline. A jury's verdict is a jury's verdict, it isn't your media plan.
- Treat automation as craft's infrastructure, not its rival. Advantage+-style systems carry the efficiency; the idea you layer on top still makes the difference.
- Labeling alone won't move Gen Z or the 65-plus audience. Both groups need a second layer of trust, usually a person or a community vouching, not a box checked by the brand itself.
Cannes gave craft a standing ovation. The balance sheet gave automation the budget. Neither one moved trust, the disclosure box did.
What this means for your next quarter
Stop reading festival applause as a budget signal. Open three separate lines: automation for efficiency, disclosure labeling for trust, craft for differentiation, and measure each against its own evidence instead of assuming one covers for the others. Heading into Q3, the question worth asking isn't "are we making a better ad." It's "are we telling the audience, plainly, what we didn't show them", because this year's data says that's the question actually buying trust.
Want this kind of thinking on your brand?
We build brand strategy, AI content and performance for the AI era.
START A PROJECT →