A new line item is appearing on publisher rate cards, and it isn’t impressions. Axios, Forbes, Future, Time and The Washington Post are packaging AI-visibility metrics into products sold to brands that want to be discoverable inside LLMs and answer engines — listicles and sponsored placements engineered to surface when a chatbot answers a question. The demand is real. The measurement is not. As Time COO Mark Howard put it about the analytics firms in this space, they “all have different methodologies. Their numbers are all different.”
01What publishers are actually selling
The product is placement inside content that AI systems are likely to retrieve, cite and surface. Forbes’ Nina Gould frames the shift plainly: “Publishers aren’t just selling impressions anymore. They’re selling visibility within the AI knowledge ecosystem.”
The context around it is the reason it has value. 56.4% of news publishers now block at least one AI crawler, which means the corpus of content AI systems can legitimately draw on is narrowing — and the publishers who remain accessible become disproportionately influential in what the machines say. Meanwhile AI-driven traffic grew 187% during 2025, a big percentage on a small base, but a direction. Time reports ranking in the 98th percentile for AI bot activity among the roughly 7,000 publishers in TollBit’s network — a vendor-supplied figure from a company with an interest in the category existing, and worth reading as directional rather than audited.
02The measurement gap is the whole problem
There is no standard for any of this. No agreed definition of an AI “impression,” no shared methodology for share-of-voice inside an answer engine, no third-party verification. Two vendors measuring the same brand’s LLM visibility will hand you two different numbers, and neither can show its work.
That vacuum has attracted what it always attracts. Insiders warn that some GEO — generative engine optimization — vendors are closer to “snakeoil salesmen” than credible partners. When the buyer can’t audit the metric, the metric becomes a sales instrument.
Publishers have found something new to sell at exactly the moment the old thing is shrinking, which is good.
03Why this matters for publishers
| You can sell this before it's standardized — but you own the credibility risk | Being early is a genuine advantage, and brands are asking. But if you sell a number you can't defend and a client later benchmarks it against another vendor's number, the discrepancy lands on your reputation, not the vendor's. |
|---|---|
| Scarcity is your leverage | With more than half of news publishers blocking at least one AI crawler, being deliberately accessible — on your terms, with telemetry — is a differentiated asset. The value of being in the corpus goes up as the corpus shrinks. |
| This is a hedge, not a replacement | AI-visibility revenue is small today against a display business under real pressure. Treat it as a new adjacency you're building competence in early, not as the thing that closes the gap. |
04What publishers should do
05The bottom line
Publishers have found something new to sell at exactly the moment the old thing is shrinking, which is good. But this is a currency with no mint. Whoever establishes credible, auditable AI-visibility measurement will set the terms for the whole category — and right now that job is open. Build your own meter, publish its limits, and you’ll be selling something a buyer can actually trust when the standards finally arrive.