Fewer than 4 percent of ChatGPT users look for news on the platform, creating a problem for media businesses looking to strike licensing deals with AI firms as AI alters digital behavior.
That low figure clashes with soaring consumer adoption. Recent industry data shows that 70 percent of Generation Alpha rely on artificial intelligence for content recommendations, while 76 percent of the general population use an AI chatbot every week.
That daily engagement generates roughly 8.2 billion AI overviews daily. Half of those citations pull from material published within the last 11 months.
Publishers are trapped. These systems rarely serve up ingested journalism as traditional news.
Instead, AI agents harvest disparate news sources to synthesize ongoing, contextual updates. They deliver actionable advice and solve immediate logistical problems for users.
Agentic AI Bypasses the Traditional Article Format
Ezra Eeman laid out the stakes at the Digital Media Europe conference in London.
WAN-IFRA’s Director of AI in Media pointed to everyday commuting as a prime example of how agentic AI operates.
An intelligent agent can track a user’s standard travel patterns, automatically scan live news reports for transit disruptions, calculate an alternate route, and execute ticket rebooking without human intervention.
The system bypasses the traditional article format entirely. It extracts raw reporting to deliver pure practical utility.
This behavior exposes a fundamental mismatch in modern publishing. Most newsrooms build editorial strategies tailored exclusively for human consumption.
That leaves them struggling to parse or index the granular information automated agents require to power complex logistics. As the defensible territory of journalism that artificial intelligence cannot replicate shrinks, Eeman warned that media organizations risk missing out on valuable licensing agreements.
Overhauling Editorial Pipelines for Triple Audiences
Dmitry Shishkin offered a different view from the same stage in London.
The independent media advisor argued that newsrooms must overhaul their editorial pipelines to accommodate multiple distinct audiences simultaneously.
Rather than focusing solely on human readers, Shishkin recommended that publishers commission and format content to serve three separate use cases:
Human audiences.
AI assistants.
Large language model crawlers.
The transition demands a complete rethink of metadata application and commissioning workflows. Original investigative reporting remains vital, but it is also costly to produce.
Shishkin noted that artificial intelligence excels at answering everyday utility-driven questions. That strength forces publishers to adapt their output.
Generative Engine Optimization Forces Technical Shifts
Generative engine optimization has already forced initial technical shifts across the industry.
Publishers now must attach author citations directly to body text. This allows automated agents to verify provenance and domain expertise.
Without clear institutional and authorial markers, agents simply bypass unverified sources.
To capture value from action-oriented agents, Shishkin advised moving away from long-form narratives.
He urged publishers to embrace single-user-need articles. Instead of publishing a single sprawling piece, outlets should break reporting into modular components designed to answer specific, isolated queries.
Reimagining the Article as an Active Agent
Eeman expanded on this modular approach.
He proposed that the article itself function as an active agent, embedding utility directly into the webpage.
In this framework, news ceases to be a destination site. It operates instead as an underlying function.
Publishers must ask what services they can integrate into external environments rather than focusing entirely on driving web traffic.
An article could dynamically explain a complex policy, monitor changing conditions over time, or compare competing options.
Shishkin observed that business-to-business media operations are already adopting these pipelines. They are building dedicated data streams for enterprise customers who require structured, agent-friendly information.
Both media specialists emphasized that these upstream structural adjustments should not replace the core mission of serving human communities.
Downstream audience engagement, regional specificity, and local trust remain essential.
By balancing utility for automated agents with distinct value for human readers, publishers can secure a sustainable foothold across both ecosystems.
