Picture the launch deck. Slide three, big type, white on black: “AI-Powered.” Everyone in the conference room nods, because the board likes it, the investors like it, and the product team spent eighteen months building the thing. Then the campaign goes live, and the customer reads those two words the way you read “contains sulfites.” Noted. Slightly concerning. Moving on.
We need to talk about the most expensive adjective in marketing.
Marketing Brew asked a handful of senior strategists this month how you advertise AI to a public that increasingly doesn’t trust it. The answers were bleak in a way that should be useful. Nika Rastakhiz, head of strategy at Mother LA, the shop behind Anthropic’s Super Bowl campaign, said that if she had to put numbers on it, at least 90% of the public conversation around AI carries a little bit of an “ick.” The remaining 10% is practical. Her team tracks sentiment on an internal dashboard, and the best you can hope for, she said, is “neutrality into slight hints of positivity.”
That’s the aspiration now. Not love. Not desire. A couple of notches above apathy.
Howard Laubscher, chief strategy officer at VML Kansas City, was even blunter: “The general term ‘AI’—just its connotation—is completely negative.” His advice, per the piece, is to sell outcomes and let the technology stay offstage.
So why does every second product page in America still lead with the word?
Because it’s easy. Because the founder wants credit for the hard engineering. Because the category got crowded and “AI-powered” felt like a flag to plant. And because for about two years, it worked. Investors heard it and leaned in. Trade press heard it and wrote it up. Somewhere along the way, a lot of brands confused the people who fund them with the people who buy from them.
Those are different audiences. One of them wants to hear about your model. The other wants to know if their Tuesday gets easier.
Dentsu’s global chief strategy officer, Patricia McDonald, put the whole problem into nine words: “Just because you build it doesn’t mean people want it.” Her team’s research, as Marketing Brew describes it, shows the skepticism isn’t even uniform. Older people tend to use AI like a replacement search engine. Younger people are more likely to lean on it for emotional support. Women are less trusting than men about handing tasks to it across every use case Dentsu looked at. So when a brand slaps one blanket “AI-powered” message on a campaign, it’s not speaking to a market. It’s speaking to an imaginary average person who doesn’t exist and wouldn’t like you if they did.
McDonald also flagged something sneakier. The pitch that sounds safest, “let AI handle the tedious stuff,” can backfire when it touches a human moment. Her example was holiday shopping. Telling someone a machine will pick their mother’s gift isn’t convenience. It’s an insult wearing a bow.
There is a version of this argument that goes too far, so let’s deal with it.
The objection goes like this: “Fine, but we are an AI company. That’s the product. Are we supposed to hide it?”
No. Hiding it would be worse. People find out, and then the story becomes that you hid it, which is the kind of reveal nobody recovers from in a single news cycle. Just look at how fast commenters now accuse handmade work of being machine-made. Get caught concealing the machine and that suspicion hardens into a verdict.
The answer isn’t concealment. It’s hierarchy.
The headline is what changes for the customer. The body copy is how. The technology is the footnote, the FAQ, the honest sentence on the product page that says what the tool does and what a human still does. The order matters. Lead with the benefit, and “AI” becomes an explanation. Lead with “AI,” and every benefit after it gets read through the ick.
Laubscher made a related point that deserves more attention than it got: people seem more willing to accept AI for logic and reasoning than for empathy and creativity. That’s a map, not just a mood. If your tool reconciles invoices, routes deliveries or flags a scheduling conflict, the machine is a reasonable thing to mention. If your tool writes a condolence card, composes a wedding toast or “personalizes” a skincare consultation with a fake warmth, keep the machinery out of the copy, and think hard about whether the product should exist in that form at all.
There’s an old precedent for this in the article that we keep thinking about. In 2007, JetBlue rolled out a customer bill of rights. It wasn’t a feature announcement. It was a promise about how you’d be treated when things went wrong. Laubscher expects AI-heavy brands to start making similar public commitments, to self-regulate loudly. We’d take it further. The brands that win the next two years won’t be the ones that say “AI” the most. They’ll be the ones that say, clearly and in plain English, what they will never let the AI do to you.
That is a positioning territory sitting wide open, and almost nobody is standing in it.
Here’s what makes this uncomfortable for our side of the business. Agencies have been just as guilty. We put “AI-enabled creative” on credentials decks because procurement asked for it, then wondered why the work started to look like everyone else’s. Rastakhiz said it has gotten a little “trope-y” for brands to lean into AI’s humanity, and she’s right. The ad where the robot learns to love is already a cliché, and the year isn’t even over.
So what do you actually write instead?
You write the outcome in the customer’s language. Not “AI-powered scheduling,” but “you’ll never double-book a client again.” Not “intelligent skin analysis,” but “find out what your skin needs before you spend another dollar guessing.” Not “AI-driven insights,” but the insight itself, stated plainly, with a number attached. If the claim is true, it doesn’t need a buzzword holding its hand. If the claim isn’t true, the buzzword won’t save it either. It will just make the eventual pile-on more fun for everyone else.
Think about how this plays on the ground. A Westside dental practice doesn’t need to tell you its scheduling runs on a model. It needs to tell you the reminder text arrives the night before, in plain English, and that rescheduling takes one tap. A Silver Lake wine bar doesn’t need “AI-curated pairings” on the menu. It needs a server who can say why this bottle works with that plate, and maybe a tool in the back that helps her remember. The customer feels the outcome. The machinery stays in the kitchen, where machinery belongs.
And when you do name the technology, name it like an adult. One sentence. What it does, what it doesn’t, who checks it. That honest sentence reads as confidence. The breathless version reads as a company that’s afraid you’ll notice it’s not that different from the last one.
Notice who’s making this argument, too. It isn’t AI skeptics on the sidelines. Mother LA made an AI company’s Super Bowl ads. The other two strategists work at some of the biggest networks in the business. The people closest to the technology are the ones telling brands to stop shouting about it. That should end the debate in most conference rooms.
It won’t, because the board still likes slide three.
So here’s our challenge for whoever is writing your next homepage. Delete the word “AI” from the headline. Read what’s left.
If there’s nothing there, the problem was never the adjective. It was that you didn’t have a reason for anyone to give a damn.