Brands Are Now Buying Ads That Humans Never See.

TIME has started selling a new kind of advertisement: one designed for software instead of people.

A normal visitor to a TIME article sees the usual webpage. Some AI crawlers can receive a stripped-down, machine-readable version containing sponsored information written for systems that power AI search and assistants. Ally Bank and the Project Management Institute were among the first advertisers testing the format.

Then came the twist: Perplexity moved to prevent those ads from influencing its systems. For marketers, the experiment exposes a question that barely existed two years ago: what happens when part of the audience interpreting your brand is a machine? (Sources: Digiday, July 30, 2026; Business Insider, August 2026.)

A New Audience Has Arrived Without a Pair of Eyes

Digital advertising has traditionally assumed that somewhere at the end of the transaction sits a human being. A person sees the search ad, watches the video, opens the email, notices the sponsored post, and decides whether to continue.

AI changes the sequence. A customer can now ask an assistant to compare banks, recommend accounting software, shortlist hotels, research suppliers, or identify the best product for a specific problem. The system may examine publisher articles, business websites, product pages, reviews, structured data, and other sources before the customer ever reaches a company’s homepage.

TIME has been adapting to that environment by producing markdown versions of webpages: simplified, text-focused copies that remove much of the visual infrastructure intended for human browsers. According to Digiday, TIME is working with ad-tech company Mobian to place sponsored FAQ-style material inside some of those machine-readable pages. The Register later reproduced examples of that bot-targeted material independently. (Sources: Digiday; The Register.)

The Advertisement Looks More Like an Answer

That format is interesting because it barely resembles traditional advertising.

Human advertising has spent decades becoming more visual: photography, film, celebrity, animation, color, typography, music, and spectacle. Advertising aimed at an AI crawler moves almost in the opposite direction. Its value lies in whether the information can be extracted, interpreted, and connected to a user’s request.

What does the company offer? Who is it for? What differentiates it? What does it cost? Where does it operate? What evidence supports the claim?

Those questions sound less like a campaign brief and more like the information architecture underneath a business.

The Register found sponsored FAQ-style material in machine-readable TIME pages delivered to selected crawlers while ordinary readers saw the normal publication experience. The commercial premise is obvious: if AI systems increasingly mediate discovery, advertisers will want their information present where those systems gather context. What remains unproven is whether placement actually changes an AI assistant’s recommendation. (Source: The Register, August 2026.)

Then Perplexity Drew a Line

The experiment immediately created a problem.

If a publisher gives an AI crawler sponsored information that ordinary readers do not receive, is that simply a new advertising format—or an attempt to influence the information environment from which an AI answer is produced?

Perplexity chose the more skeptical interpretation. Business Insider reported that the company moved to stop TIME’s bot-targeted advertisements from influencing its systems and compared the practice with “cloaking,” the long-running search tactic of serving different material to crawlers and users. Perplexity also indicated that publishers using deceptive approaches could risk being downranked. (Source: Business Insider, August 2026.)

That response matters because it turns a clever media experiment into a governance problem. The emerging AI economy now needs to distinguish useful commercial information from paid attempts to manipulate machine judgment.

Search engines have seen versions of this fight before. Keyword stuffing, hidden text, link schemes, and crawler-specific pages all emerged as businesses tried to influence ranking systems. Search platforms responded with policies designed to separate relevance from manipulation.

AI assistants may now be entering their own version of that contest.

Every Business Now Has Two Readers

Even companies that never buy an “AI agent ad” should pay attention.

A business website increasingly serves two audiences. The first is familiar: the person. That person needs clarity, confidence, proof, relevance, and a reason to act.

The second is the system interpreting the business for that person. That system needs information it can identify, compare, and reconcile.

A restaurant can describe itself beautifully while making its actual hours, menu, location, or reservation rules difficult to determine. A software company can publish hundreds of articles while leaving its pricing, capabilities, and ideal customer ambiguous.

A professional-services firm can sound sophisticated while never stating plainly what problem it solves.

In conventional marketing, those failures create human confusion. In AI-mediated discovery, they can create machine confusion too.

This is where the OrionPilot point of view becomes useful: strategy, positioning, content, channel execution, and evidence should describe the same business consistently.

The answer is not to create a second brand identity “for AI.” It is to make the existing one coherent enough that both people and systems can understand what the company actually does.

Marketing Information Is Becoming Infrastructure

For years, marketers asked: can customers find us?

Social platforms added another question: can we earn their attention?

AI introduces a third: can the systems helping customers make decisions understand us correctly?

That does not mean writing websites for robots. It means recognizing that marketing information increasingly functions as infrastructure.

Positioning is information. Pricing is information. Product descriptions are information. Reviews are information. Policies, locations, availability, evidence, expertise, and customer outcomes are information.

An AI system may encounter all of them before the customer encounters the homepage.

TIME’s experiment may become an important advertising format or disappear as AI platforms establish stronger boundaries. That outcome is still unknown. What is already clear is that advertisers are testing ways to reach machines as intermediaries, while AI companies are deciding what kinds of commercial influence they will tolerate. (Sources: Digiday; Business Insider.)

The competitive question is therefore getting larger than visibility.

Businesses still need to be noticed by people. But increasingly they may also need to be accurately understood by the systems people ask for help.

For the first time in digital marketing, a meaningful part of the customer journey can begin with an audience that does not watch, scroll, or click—yet may influence what a human considers next.