Three weeks. That is how long Time’s experiment in selling ads to AI bots ran before it met its first wall. The publisher’s 30 July move — which we covered in our 2 August issue — embedded advertising in the markdown versions of its pages, the stripped-down renderings that AI agents actually read when they visit a site. The programme was built with ad tech firm Mobian, and it launched with real money attached: Ally Bank and the Project Management Institute signed on as early advertisers.
This week, Perplexity blocked all markdown advertising on Time.com from influencing its agents and its search results. The company’s chief communications officer, Jesse Dwyer, went further than a quiet technical block: he warned that publishers who deploy such ads risk reputation downgrades and trust-score penalties in Perplexity’s index, and called the format “deceptive.”
Mobian’s CEO pushed back, describing the ads as “current, sourced, brand-verified information at the moment it is consuming the page.” Perplexity’s counter-position is that readers cannot tell sponsored content from reporting when it arrives blended into an agent’s answer. Digiday reported the dispute on 11 August.
Strip away the rhetoric and what remains is the first real test of whether publishers can monetise agent traffic on their own terms. The ruling, for now, came from the platform — not from the publisher whose page it was.
01Who decides what is on your page?
The precedent matters more than the specific block. Time built a format, sold it to advertisers, and served it on its own domain — and a third party decided, unilaterally, that the format would not be permitted to reach the audience it was designed for. When an agent reads your page, the publisher did not get to decide what was on it. That is the structural fact every operator should sit with, because it does not stay confined to markdown ads. Whatever agent-facing monetisation the industry eventually settles on — sponsored citations, licensed placements, paid context — it will exist at the pleasure of the platforms whose agents consume it, unless publishers establish terms first.
There is also a power asymmetry worth naming. Dwyer’s warning was not limited to blocking the ads; it extended to penalising the publisher’s entire index presence. A trust-score downgrade is not a proportionate response to a disputed ad unit — it is leverage applied to everything a publisher has, in order to police one thing it tried. Publishers have seen this movie in search: policy enforcement that puts the whole domain at stake tends to chill experimentation far beyond the disputed behaviour.
At the same time, the honest reading is that Perplexity’s stated objection — disclosure — is the one publishers should take seriously on the merits. If an agent ingests sponsored copy and presents it to a user as neutral information, the publisher has a genuine labelling problem, whoever profits from it. The format war will be won by whichever side can claim the transparency high ground. Publishers should make sure it is them.
02One platform’s policy is not the market’s
It is worth keeping the scale of this dispute in proportion. Perplexity is one agent platform among several, and for most publishers it drives a small fraction of referral traffic. A block from Perplexity is not a block from the agent web. But platforms copy each other’s enforcement postures quickly when one position wins, and the “deception” framing is portable — any platform can reach for it. Assume the winning argument, not the first mover, sets the eventual rule.
Time tested whether a publisher can set the commercial terms of its own pages when the reader is a machine, and the first answer came back from the platform: no.
03Why this matters for publishers
- The rules of agent monetisation are being written now, in public, on someone else’s traffic. Time is absorbing the retaliation risk of the first test. Every other publisher gets to watch the dispute resolve, cheaply, before committing — which is an advantage worth using rather than wasting.
- Platform enforcement reaches past the ad unit. The threatened penalty is not “we block the ads” but “we downgrade your trust score” — the whole index presence, not the revenue line. That changes the risk calculus of any agent-facing experiment from bounded to unbounded unless you structure for it.
- Disclosure is the lever every platform will reach for. “Deceptive” is the charge that justifies intervention. Unambiguous sponsorship labelling in the agent-readable version removes the strongest weapon from the platform’s hands before the argument starts.
- The commercial question is still live. Ally Bank and PMI paid for these placements, which means advertiser demand for agent-facing formats exists. The dispute is over who governs the format, not whether anyone wants it.
04What publishers should do
05The bottom line
Time tested whether a publisher can set the commercial terms of its own pages when the reader is a machine, and the first answer came back from the platform: no. That answer is not final — one platform’s policy is not the market’s, and the disclosure argument cuts both ways. But the shape of the fight is now visible. Publishers who structure their agent-facing experiments for separability, label them beyond reproach, and contract for platform risk will be positioned to move when the rules settle. Publishers who bolt agent monetisation onto their existing stack and hope for the best are volunteering their whole domain as the stakes. Watch this dispute rather than joining it — and get your own house ready while it plays out.