For publishers across continents, the question is no longer whether AI will reshape journalism. It is how news organizations should engage with it — defensively, commercially, editorially or as a new audience in its own right.
The sessions revealed a global industry testing sharply different strategies.
The Hindu Group in India is using AI to solve a discovery problem. Bonnier News in Sweden is applying AI to personalization and conversational archives. Kleine Zeitung in Austria is rebuilding its publishing strategy around the idea that AI bots are becoming a new class of readers.
TollBit is positioning itself between publishers and AI companies to track scraping, licensing and future agentic use.
Together, they pointed to a new phase in the AI transition: from using AI inside newsrooms to designing journalism for an AI-mediated internet.
For Pundi Srirami, chief product officer and business head of STEP at The Hindu Group, AI is not primarily a replacement for journalism. It is a way to help readers find journalism they might otherwise miss.
"Our sites are full of stories our subscribers never saw," he said.
The Hindu reaches about 35 million monthly users, with roughly 1 million users in its app. But converting reach into paying relationships remains difficult. Srirami said the challenge is not only willingness to pay, but discovery: how to expose readers to the right story, in the right format, at the right moment.
The Hindu is using AI to repackage the same story into multiple forms. Readers can choose AI summaries, Q&A formats, short versions of around 200 words, or longer versions of around 300 words. Engagement with these AI-generated formats rose from 6% to 36%, according to Srirami.
The company is also personalizing not just content, but the surfaces on which stories appear. AI-driven trending modules, push notifications and personalized app areas are now contributing meaningfully to app usage, with personalized surfaces accounting for about 15% of app page views and reaching around 30% in some areas.
Audio is being used selectively. The Hindu has found value in audio explainers for dense daily editorials, particularly for readers preparing for state examinations. Those explainers have reached a 24% completion rate.
Srirami said the AI-driven features are now contributing real business value, accounting for roughly half of app engagement growth over the past 12 months.
At Sweden's Bonnier News, AI is being used to make personalization more precise and archives more accessible.
Jan Helin, chief product officer at Bonnier News, said AI allows publishers to move beyond broad recommendation systems toward more accurate, direct reader interactions. One promising area is conversational access to archives, where readers can ask questions and receive answers drawn from the publisher's own journalism.
The shift is from search to conversation. Instead of expecting users to know the right keyword, publishers can let them ask what they want to know.
Helin said reader requests through conversational interfaces are producing strong results, with conversion rates reaching around 60%.
The most aggressive vision came from Sebastian Krause, head of digital at Austria's Kleine Zeitung.
Krause argued that publishers spent the past 15 years optimizing for Google, chasing clicks and attention. Now, a new kind of visitor is emerging: AI bots.
Krause's conclusion was blunt: publishers should stop thinking only about human traffic and start treating AI as a new audience.
"Don't bet against a great product," he said. "AI is a great product."
For Kleine Zeitung, that means rethinking the newsroom and the product architecture. Krause described the possibility of one site for humans and another built for agents: machine-readable, accessible and designed for AI consumption, but on terms set by the publisher.
The goal is not to give everything away. It is to protect content value, monetize AI usage and dictate the conditions of access.
"They can summarize or pay," he said.
Krause argued that retrieval-augmented generation, or RAG, may be a larger and more sustainable opportunity for publishers than training data. AI systems will need reliable, high-quality sources to answer current and specific questions. That could make publishers valuable not because they produce general news, but because they produce what only they can report.
In the AI era, he said, distribution is no longer the main bottleneck. Originality and scarcity are.
For publishers, the challenge is to build systems that make AI access measurable and monetizable rather than invisible and uncompensated.
Copyright ⓒ Aju Press All rights reserved.
