Gartner's new CMO Spend Survey landed this week, and the headline isn't a tool — it's a tell. Two-thirds of media dollars now chase conversion and awareness, and loyalty has fallen below 15 cents of every media dollar. AI optimizes toward whatever it can measure, and the easy-to-measure bottom of the funnel is winning. But here's the good news the data hides: the most AI-ready teams spend more on loyalty, not less. The opportunity is to let AI grind the optimization and reinvest the time it frees up into the brand and the customer relationships a machine can't fake.
What happened. Gartner's 2026 CMO Spend Survey — 401 CMOs, fielded January through March — dropped fresh numbers this week. As Marketing Dive reported, 62.6% of media spend now goes to conversion and awareness, up 10% from 2024. Loyalty and retention have slipped under 15% — a 29% drop over the same stretch. And even as CMOs chase AI to cut costs, they're spending more on people: labor rose to 24.5% of the marketing budget, up from 21.9% a year ago.
Why it matters. AI is brilliant at optimizing whatever you can measure, so the dollars drift to the click you can track today instead of the customer you keep for five years. That's the old brand-versus-performance seesaw — and AI just put a thumb on the short-term side. The fix isn't to fear the technology; it's to point it. Gartner's own data shows the tell: the most AI-mature teams invest more in loyalty and retention, not less. Maturity isn't using AI to tune the funnel faster — it's using the hours AI gives back to protect long-term value. And notice the humans: CMOs are hiring people while they adopt agents. AI isn't thinning the team; it's raising the price of judgment. As Gartner's Ewan McIntyre put it, "optimization is not the same as strategy."
A new MediaScience and Ehrenberg-Bass study of 900 viewers found AI disclosure labels didn't dent brand recall, recognition, or attitude. A clear text label actually nudged unaided recall from 54% to 61%. New York's synthetic-performer law took effect in early June.
What to do: add a plain-text AI disclosure to one AI-made asset and stop treating transparency like a tax.On June 11, Mediaocean launched NIVO, an AI intelligence layer, and Magnite shipped orchestration that lets buy-side AI agents transact directly with premium supply. The agentic shift is moving past the walled gardens into the open pipes.
What to do: ask your media partners which of your buys an agent will touch in the next six months — before one already has.GSTV, the national video network at fuel pumps, is stepping up audience targeting with Stagwell's agentic system — proof that agents aren't just buying media, they're building the segments behind it, well beyond your social feed.
What to do: name one channel where you'd let an agent build the audience, then watch it for a quarter before you trust it.Geico put its Gecko in a chair it had never sat in: a live, unscripted conversation. On June 11 the mascot called into iHeartMedia's "Fudd Around and Find Out," the podcast Geico sponsors, and traded real-time banter with WNBA star Azzi Fudd — powered by AI, with The Martin Agency and visual-effects house Framestore behind it. The guardrails are the story: humans reviewed the Gecko's real-time output, and the activation stayed inside SAG-AFTRA guidelines. New CMO Arianna Orpello framed it plainly — "Geico does its best work when we don't follow the category, we rewrite it." Coca-Cola is running a similar play, pitting soccer manager José Mourinho against a digital twin of himself for the World Cup.
Three concrete moves a marketer could make. If you do nothing else from this issue, consider these.
Pull your funnel split — how much goes to conversion and awareness versus loyalty and retention. If AI tools have quietly tilted you toward the bottom of the funnel, you'll see it in the numbers. Name it before the next budget review does. By Friday.
Pick an ad or post where you used AI and add a clear text disclosure. The new data says it won't cost you recall and may even help — and with state laws landing, you'd rather practice now than scramble later. Before your next launch.
On June 12 the U.S. government ordered Anthropic to pull Claude Fable 5 and Mythos 5 — and because the lab couldn't wall off foreign users, it had to shut both down for everyone, worldwide. Is this the first step toward standing oversight of frontier models, or a one-off national-security order? Nobody knows yet — so don't hard-wire one model into a workflow you can't run without. List your AI-dependent workflows and name a fallback for each. By next issue.
This is not an assignment. It's for your use as you see fit — read it, skip it, forward it, file it. The Playbook is a set of options, not a checklist.
Michigan sentiment ticked up to 48.9 in June but is still 19% below a year ago, while prices climbed 4.2% — the hottest in nearly three years — the shopper your AI is optimizing toward is paying more and trusting less, and that's where your empathy earns its keep.
AI in Marketing Weekly is written by Ged King alongside the Master of Digital Marketing & AI — a program where faculty and students alike are working professionals, applying this in real jobs every week, not studying it from the sidelines.
Explore the program →