News companies must extend audience funnels to machines
Conference Blog | 12 March 2026
News publishers must rethink audience funnels for a world where AI systems increasingly consume journalism, Greg Piechota told media executives Thursday during the INMA Media Subscriptions Summit in Toronto.
Piechota, researcher-in-residence, INMA Readers First Initiative lead, and summit curator, argued that the next phase of digital transformation will require publishers to treat machines as audiences and customers — not just humans.
“In the last year for the first time ever, most of your online visitors are machines. They’re no longer humans,” Piechota told the 223 summit attendees from 23 countries.
As AI reshapes discovery, publishers must expand traditional audience funnels to include the systems that crawl, evaluate, and synthesise journalism, he said.
Discovery is fragmented, not broken
The traditional linear funnel — awareness, engagement, subscription — no longer captures how people discover news, Piechota said: “The funnel didn’t break; it just fragmented,” he said.
Instead, audiences may encounter journalism through social media, search, individual journalists, or newsletters before eventually forming habits that lead to subscriptions.
In that environment, discovery should be reframed as a pipeline for long-term value rather than free distribution.

“Those who grew subscriptions last year had a healthy top of the funnel,” Piechota said. “We need to have it healthy top of the funnel, we just need to call it differently. This is a pipeline for your customer lifetime value.”
The early stages of the subscriber journey are especially critical.
“Most churn actually happens within the first week, the first month, the first three months. By the fourth month, 50% of new subscribers usually leave,” he said.
That insight, he said, means publishers should prioritise onboarding and engagement rather than focusing resources later in the lifecycle.
Machines are already part of the audience
News publishers are not new to serving machines. Search engines have long evaluated journalism through technical signals, authority, and user behaviour, Piechota said.
AI systems extend that process by performing multiple searches, gathering information from various sources, and synthesising answers.

“They will look probably for the signals of authority. They will look for signals for trust,” he said.
What has changed is how people interact with these systems.
“Most queries to Google are very informational. Most queries to ChatGPT are actually action-based,” he said. “We don’t ask ChatGPT mostly ‘tell me about’ — we ask it to write the f***ing memo.”
Despite that shift, AI systems still rely on credible information grounded in verified reporting.
Case study shows how SEO is evolving for machines
The shift toward machine audiences was also reflected in a case study shared during the summit by Shelby Blackley and Jessie Willms of WTF is SEO, who outlined how search behaviour is changing as AI systems increasingly mediate discovery.
Their work highlighted how traditional SEO assumptions — optimising for clicks and human readers — are being challenged by AI-driven interfaces that summarise content and reduce direct traffic to publishers’ sites.
Instead, publishers must ensure their journalism is structured, authoritative, and machine-readable so it can be surfaced, interpreted, and prioritised by AI systems.
The case reinforced Piechota’s argument that publishers are no longer optimising discovery solely for human users but also for machines that decide which content is seen — and how it is presented.
Finding ways to charge AI platforms
As AI companies increasingly use news content to generate responses, Piechota said publishers must develop monetisation strategies that capture value from machine audiences.
Possible approaches include micropayments for crawling or content usage, extending subscription infrastructure to machines, or licensing journalism to enterprise platforms. Another strategy is focusing on selling full archives, datasets, or feeds rather than individual pieces of content.
“We need to focus on selling ongoing access to full archives, datasets, or feeds,” he said.
Selling individual facts or fragments, he warned, risks undermining the value of journalism.
Strategy matters more than technology
Digital disruption historically has been driven by customer behaviour more than technology, Piechota stressed: “In digital disruption, it’s customers who disrupt companies, not just technology.”
For publishers responding to AI, the priority should therefore be strategic thinking about business models.
“We need a business strategy for the world changed by AI much more than we need an internal AI strategy,” Piechota said.
Scarcity in an age of infinite content
As AI accelerates the production of information, Piechota argued the most valuable qualities in journalism will be those machines cannot easily replicate.
“Attention is actually scarce,” he said. “Human judgment is scarce. And accountability for the judgment.”
Journalists’ responsibility for the information they publish — and the trust that creates — may become one of the most valuable assets for news organisations in an AI-driven ecosystem.
Piechota ended with a straightforward message for publishers navigating the shift: “Treat the machines as customers,” he said. “Make them pay.”
The summit ends today with topic-specific seminars.
Photos by Robert Downs Photography.








