Publishers Pitch AI Visibility to Advertisers as the New Comscore

Time fielded several advertiser requests last week, each one asking how to get chatbots talking about their brand. Time already had a pitch ready to sell them.

Axios, Forbes, Future, The Washington Post and a cluster of German publishers are running the same play, packaging their prominence inside ChatGPT, Google’s AI Overviews and Perplexity, all built on large language models (LLMs), into a new number they can sell straight to advertisers: AI visibility. Nobody agrees on how to count it, and the crawlers producing those citations take far more from publisher sites than they hand back.

Axios and Time Turn Citations Into a Sales Pitch

Publishing executives say the shift did not start inside their own marketing departments. Brands came to them first, asking how to read their footprint inside AI answer engines.

Time’s chief operating officer, Mark Howard, said the magazine fielded multiple requests for proposals last week alone, all seeking ways to improve how a brand shows up in AI platforms. Axios pushed the idea into an actual product. After ranking among the most-cited publishers in AI answer engines across four separate studies, including ones from Muck Rack and 5W, the company saw an opening to sell that standing, said chief revenue officer Jacquelyn Cameron.

Cameron said the metric carries real weight for the brand: “Ensuring that Axios has a high citation score within these LLMs is something that we think about a lot… That’s incredibly important for us as a brand and as a news publication, especially as adoption of these tools continues to accelerate.” Axios is already holding early conversations with advertising clients about selling that same insight, she said.

Forbes is pitching something similar. Chief innovation officer Nina Gould said the value proposition is evolving beyond audience reach, engagement and SEO to include how authoritative content from trusted media brands shapes how AI systems understand, cite and recommend brands over time.

A wave of vendors wants a piece of the same trade. Content distribution platform Stacker tracks story performance across six AI platforms, a feature it launched in late June that tests each piece of content against 30 prompts over four weeks. It joined a crowded field of tools all claiming to measure the same murky thing.

Why Referral Traffic Cratered First

The push for a new metric follows a real financial hole. Publisher ad supply fell by up to 40% in the second quarter, according to benchmarking data from ad marketplace Ozone, as zero-click search choked the flow of traffic to news sites across the US and UK.

Google search traffic to publishers dropped by a third worldwide in the year to November 2025, according to Chartbeat data cited in the Reuters Institute’s 2026 trends report. Lifestyle, health and travel sites, the categories that leaned hardest on search-friendly how-to content, have felt it worst.

None of that decline has been offset by traffic AI platforms send back. AI-driven traffic to the open web grew 187% in 2025, about eight times faster than human traffic growth, according to web data platform Decodo. Growth that fast still starts from a tiny base.

Nobody Agrees on How to Measure It

Proving a publisher is a heavily cited source inside AI answer engines is the strongest signal of authority available right now, one publishing executive said on condition of anonymity. Brands report that their own products keep surfacing whenever they ask a chatbot about them, the executive added.

The industry has taken to calling this pursuit Generative Engine Optimization, or GEO, the practice of shaping content so AI systems select and cite it. No two trackers count it the same way. Most analytics firms measure how often a publisher is mentioned or cited in AI answer engines, but they pull from different datasets and different methods, and their rankings rarely line up. Forbes uses a platform called Profound to track citations, share of voice, source attribution and competitive rankings across thousands of prompts, Gould said.

Howard put it more bluntly.

They all have different methodologies. Their numbers are all different. I don’t think that there’s any standardization here that you can anchor on, at the moment.

Howard said the industry has nothing close to the common yardstick Comscore or Similarweb once gave the open web. Time instead leans on bot traffic as its main proxy with clients, tracking daily and monthly AI crawler activity against human visits and comparing citations with pageviews.

The Crawl Ratio That Undercuts the Pitch

Even if publishers agreed on how to score their AI visibility, the economics sitting underneath it are lopsided. Automated requests overtook human traffic on the open web for the first time in June, with bots crossing 57.5% of HTML web traffic against 42.5% from people, Cloudflare chief executive Matthew Prince said.

Training crawlers make up most of that automated activity, and they send almost nothing back to the sites they read. Cloudflare attributed 51.8% of AI crawler requests to model training in May, versus just 9.3% for search, meaning the bots most likely to cite a publisher inside a live answer are a small slice of total AI traffic.

The gap between platforms is stark, and it swings fast month to month. A crawl-to-referral ratio tracker updated weekly, built on Cloudflare Radar data for the week spanning May 25 to June 1, puts Anthropic’s Claude crawler at the bottom of the reciprocity ladder.

Every ratio in that table moved during the prior month alone, and only Google’s held roughly steady. Perplexity’s got worse as its crawl volume outran what it sent back. Google’s traditional search crawler, the model publishers built an entire industry around serving, still returns a visitor for close to every five pages it reads.

Publishers have responded by shutting doors rather than waiting for the ratio to improve. According to web data firm HasData, 56.4% of news publishers now block at least one AI crawler through their robots.txt files. A BuzzStream survey of top publishers found a separate figure closer to 79% specifically for blocking AI training bots, a reminder that even blocking rates depend on whose count you trust.

Is Anyone Buying GEO Yet?

Not many, by publishers’ own account. Executives describe early conversations rather than signed contracts, and several of the industry’s biggest names say the entire premise is unproven, since no vendor has shown a repeatable way to influence what a chatbot actually says.

Where the industry splits on GEO’s value:

• Axios says the metric already drives real business conversations. Cameron said that regardless of which analytics company a brand asks, “we are still populating as one of the top sources.”
• Time treats vendor rankings with open skepticism but leans on its own bot-traffic data instead. Howard said Time compares AI crawler visits against human traffic to make its case to advertisers.
• People Inc. rejects the whole premise. Chief executive Neil Vogel said the entire conversation is not rooted in fact, and challenged anyone claiming they can optimize chatbot output to prove it.

Vogel’s skepticism is not new. He made the same case to Digiday in 2025: “This was all based on the assumption that somebody out there knows how to optimize for this. This whole conversation is not rooted in any fact. If there’s anyone who can prove to me that they can optimize the output of these rapidly developing tools, I would love to talk to them.” An SEO manager at a large news publisher offered a milder version of the same doubt, speaking on condition of anonymity: “I’m a little skeptical of anyone who is promising results at the moment.”

Publishers Are Building Their Own Scoreboard Instead

A separate, more structural effort is underway alongside the ad-sales pitch. A publisher-led coalition called SPUR, short for Standards for Publisher Usage Rights, wants to replace opaque scraping with a shared system for tracking exactly how AI companies use publisher journalism.

Founded by the BBC, the Financial Times, Guardian Media Group, Sky News and Telegraph Media Group, the coalition has grown quickly. The Associated Press joined the coalition earlier this month, pushing membership past 30 publishers and six affiliates. Its Content Telemetry standard breaks AI content use into five measurable events:

• Retrieval, when an AI system pulls a publisher’s content, something sites can already detect.
• Grounding, when that content gets loaded into an AI application to help generate a response.
• Citation, when the content is explicitly named as a source in an AI answer.
• Display, when the content is shown or referred to inside the response itself.
• Engagement, when a user actually acts on or clicks through to the underlying publisher content.

The coalition opened the draft standard for public comment in June, with a London event that drew Microsoft and content-delivery network Fastly into the room. Licensing and infrastructure startups including TollBit, Redpine and MonetizationOS have said they plan to build the standard into their own products.

The hard part is the same one that has dogged every publisher alliance before it. SPUR can define how AI usage should be measured, but it cannot force OpenAI, Google or any other model maker to report it. The logic, as Digiday has reported it, is that no single publisher, even a large one, can force an AI company to accept publisher-friendly terms on its own, but a large enough group moving together might.

Scott Messer, principal of Messer Media, has said the effort only works if standards and enforcement are treated as a shared project across the industry. Publisher alliances have a mixed record on that front. During the rise of programmatic advertising, shared industry systems often created value for platforms while leaving publishers with less control over their own inventory.

Public comment on the telemetry draft runs through July 24. Whether any AI company agrees to use it is still an open question.

Frequently Asked Questions

What Does AI Visibility Mean for a News Publisher?

It measures how often, and how prominently, a publisher’s journalism gets mentioned, cited or used to ground an answer inside a chatbot or AI search summary, rather than how many clicks or pageviews that journalism receives. Publishers are now packaging that footprint into products they sell directly to brands.

How Is Time Measuring Its Own AI Visibility?

Time tracks daily and monthly AI bot traffic against human visits and compares its AI citations with pageviews. Chief operating officer Mark Howard has said Time ranks in the 98th percentile for AI bot activity among nearly 7,000 publisher sites tracked by licensing platform TollBit, meaning almost no other site in that network sees more AI crawler requests.

What Is the SPUR Coalition’s Telemetry Timeline?

SPUR published its draft Content Telemetry standard in June and opened it for public comment through July 24, 2026. The coalition, whose founding members include the BBC, the Financial Times, Guardian Media Group, Sky News and Telegraph Media Group, now counts The Associated Press among more than 30 publisher members and six affiliates.

Is AI-Referred Traffic Replacing What Publishers Have Lost?

Not yet, at scale. AI-referred traffic to commercial sites jumped 632% year over year, but it still made up just 0.2% of total web traffic, according to analytics firm Contentsquare. The growth rate is dramatic, but the base it is growing from remains tiny next to what search referrals used to deliver.

Do Brands Care About Showing Up in Chatbot Answers?

Marketing agencies say yes, and increasingly so. Ashish Jacob, director of GEO at digital marketing agency Go Fish Digital, said visibility in AI responses is becoming an important measure of influence, even when traditional traffic and attribution numbers do not fully capture its impact.