Interested in talking?
|
          AI Search Visibility

          Ad AgeThe Year Brands Learned What AI Really Thinks

          The most important marketing conversations I had this year weren’t with people. They were with machines.


          The most important marketing conversations I had this year weren’t with people. They were with machines.
           

          That realization crept up on me slowly, then all at once. This was year three of the AI revolution, but it was the first year it stopped feeling experimental and started feeling structural. AI wasn’t a tool anymore. It was the room.

          It seeped into education, work, travel, finance, law, parenting and—inevitably—my already fragile attention span. Markets rode the wave like a roller coaster sponsored by five mega-cap stocks. Underneath the froth were harder questions marketers are still struggling to answer: What does learning look like now? Where does trust live? And who—or what—decides when an answer is “good enough”?

          By midyear, one thing became clear: this wasn’t a content arms race. It was an answer credibility crisis.

          When text lost its monopoly

          One of the biggest shifts arrived quietly: multimodal finally became real.

          Brands aren’t just being described anymore. They’re being visualized, simulated and demonstrated. Image generation models advanced to the point where they could produce accurate composition, readable text and fewer hallucinations.

          That raised the stakes. A wrong sentence was one thing. A confidently wrong image of your product, packaging or usage instructions was something else entirely. This year, brand risk went visual.

          Slop, signal and the shrinking half-life of novelty

          Like many marketers, I spent part of the year experimenting with the advancing tools—songs, parodies, cartoons, creative detours across platforms like Gemini and ChatGPT that were fun, fast and fleeting.

          The lesson wasn’t that creativity disappeared. It got cheap. Novelty decayed faster than expected. LinkedIn humbled my false sense of authentic creativity by adding their “AI Generated” watermark to many of my images. (OK, that’s fair.) What endured was signal: clarity, consistency and repeatable truth.

          The same dynamic showed up in how work got done. AI compressed weeks of production into hours. Slides became collages. Prototypes appeared overnight. Velocity surged. Judgment became the constraint.

          Creativity scaled. Discernment didn’t.

           

          Search didn’t die—it lost the moment

          Another shift landed with less fanfare but more consequence: for high-intent questions, search stopped being the center of gravity.

          Google AI Overviews, ChatGPT, Perplexity, Claude, even—dare I say—Amazon Rufus and a wee bit of Walmart Sparky became default starting points for “explain,” “compare” and “should I?” moments. With OpenAI now processing over 2.5 billion prompts per day, discovery collapsed into a single conversation. Brands began getting evaluated before anyone clicked.

          The funnel didn’t just shrink. It inverted.

          The uncomfortable truth: AI doesn’t trust marketing

          Across the year, one insight kept surfacing: AI doesn’t trust campaigns. It trusts documentation.

          FAQs. Help centers. Policy pages. Regulatory disclosures. Third-party validation.

          That reality became impossible to ignore. Working alongside close to 60 global brands trying to make sense of AI visibility, vulnerability and readiness, a clear pattern emerged. AI surfaced truths brands weren’t always eager to hear—not because it was cynical or adversarial, but because it was literal. It reflected what was documented, what was consistent and what could be corroborated.

          In that sense, it became an uncomfortable truth teller—a kind of automated BS detector, applied evenly and without apology.

          Websites quietly shifted roles, behaving less like storytelling hubs and more like evidence lockers. For many CMOs, that realization landed somewhere between unsettling and clarifying.

          Plausibility outran proof—and trust became a tax

          AI also got better at sounding right, even when it wasn’t.

          This year, plausibility began to outrun proof. Wrong answers became smoother, more confident and harder to detect. Trust stopped being a messaging problem and became an operational one. If your answers didn’t line up across systems, AI noticed.

          But the bigger consequence showed up in answer structure. Brands with trust deficits began paying what I started calling the Trust Tax—a computational penalty that showed up in how AI hedged, qualified and framed their mentions.

          Where a trusted brand got: “Brand X offers...”, a questioned brand got: “Brand Y claims... though some reviews note...” The difference wasn’t just reputational anymore. It was algorithmic. Trust became a tax rate applied to every answer, every time.

          For CMOs used to thinking about trust as a brand metric, this was clarifying: trust wasn’t just affecting perception. It was affecting computational load, answer length, and competitive alternatives surfaced.

          Consistency became visible. Gaps became scalable.

          When agents start acting

          The most profound change wasn’t smarter chat. It was delegated decision-making.

          Shopping agents compared. Travel agents booked. Research agents summarized. Software began acting on our behalf—quietly, relentlessly.

          Brands realized they weren’t just persuading people anymore. They were persuading delegates—systems with no nostalgia, no loyalty and no patience. And here’s where it got uncomfortable: private label appeared to be triumphing. Shopping agents seeking the perfect blend of performance and price were crazy kind to Kirkland. When algorithms optimize for rational value without emotional attachment, brand equity faces a different kind of test.

          Why I started writing it down

          Some days, I’m wildly optimistic. Other days, I worry the disruption will be deeper, harsher and more uneven than leaders are willing to admit.

          That tension—and the growing realization that answers, not ads, are becoming the scarcest resource—is why I spent the second half of the year writing. “The Answer Economy: How Agents, Not Ads, Shape the Future of Brands” (coming from Wiley) is my attempt to make sense of what comes next.

          Because in a world of infinite slop, answers win.

          AI will answer questions about your brand, whether you’re ready or not. It doesn’t reward spend. It rewards clarity. It doesn’t trust slogans. It trusts evidence.

          As we head into another year of acceleration, here’s the question worth asking: What would AI answer about your brand—today?

          Related posts