Nobody was watching.
Something else was already deciding.
McKinsey just put a number on it: more than half of advertisers now say AI agents already decide what gets discovered, considered, and bought — before a person ever looks. Weeks later, a regulatory clampdown made clear the recognition fight inside that economy is still wide open.
For twenty years, the internet ran on one currency: attention. Get seen, get clicked, get followed, and the rest of the business model took care of itself. On June 16, McKinsey published survey research finding that currency is already losing its job. More than half of advertisers surveyed now say AI has already reshaped how people discover and weigh what to buy — not through scrolling, but through agents that research, compare, and decide with limited human input at the moment the decision actually happens. Nobody voted on this. It became true while the market was still optimizing for eyeballs.
This is not a trend piece. It is the handoff Three Epochs of Authority™ named before McKinsey had numbers for it. Epoch 1 ran on institutional credentials. Epoch 2 ran on platform-era negotiation — followers, engagement, the attention economy McKinsey just pronounced obsolete. Epoch 3, the Inference Era, runs on what a system can verify, classify, and recommend before a person ever opens a tab. The attention economy is not dying of natural causes. It is losing to a system that was never trying to get anyone’s attention. It was trying to decide.
The audience changed. Most firms haven’t noticed yet.
The audience changed. The reader a page now has to satisfy first is not a scrolling human — it is another system, deciding on that human’s behalf.
The currency changed. Clicks and follows measured attention. Citation and classification measure whether a machine trusts you enough to act.
The decision moved upstream. By the time a person sees a result, an agent has often already narrowed the field — the moment that used to be won with attention now closes before attention is even possible.
The infrastructure fight is getting harder. The recognition fight just got wide open.
Weeks after McKinsey’s numbers landed, New York became the first state to ban new hyperscale AI data centers outright — a year-long moratorium Governor Hochul signed citing a power grid that cannot absorb the nearly 12 gigawatts of pending demand. Building the machines that run the agent economy is now a fight with governments, not just budgets. Recognition inside that economy is not — yet. The FTC’s own open comment docket on AI-accuracy claims is the same story from the other side: the rules for who a machine is allowed to trust are being written in public, right now, not handed down finished.
Answer Engine Authority™ exists for exactly this seam — the interval between a currency changing hands and the rules catching up. AEA™ does not compete for the infrastructure. It installs the entity architecture, the signal consolidation, the third-party corroboration an agent economy actually checks before it decides on someone’s behalf without asking. The firms doing that now are not early. They are on time, for a window that will not still be open a year from now.
Fifteen editions into this record, the throughline hasn’t moved: something decides who gets recognized, who gets seen, who gets cited — increasingly without a person ever being asked. McKinsey just gave that pattern a number. New York just gave the infrastructure a moratorium. The recognition fight is still nobody’s to lose. For now.
McKinsey’s February 2026 AI in Advertising Survey found that more than half of advertisers now say AI has already reshaped product discovery and consideration, with agentic purchasing decisions occurring with limited active human input — published June 16, 2026, in "The agentic advertising economy: From attention to action".
New York Governor Kathy Hochul signed a year-long moratorium on new hyperscale AI data centers on July 14, 2026, citing nearly 12 gigawatts of pending grid demand — reported by CNBC.