Based on studies conducted in Spain, Pepe Cerezo Gilarranz, founder and director of Digital Journey, shared key insights about the impact of AI the media business during a recent INMA Webinar.

During The Impact of Artificial Intelligence on Business Models, Cerezo expressed concern about the future of employment in journalism and how AI forces us to reinvent ourselves to remain relevant as content creators. 

Cerezo argued the “deconstruction” of media means building strong interest groups. 

“We have to abstract ourselves from the content, which of course remains very important,” he said, warning that if the news industry can’t attract audiences in a lasting way and build communities, its ability to survive will be limited. 

According to Cerezo, these communities rest on four pillars, including the importance of both global and local brands that reflect the company’s mission and values. 

“This is obviously a business, but it must be considered a public service,” he said, highlighting journalism’s societal role.

This mission of service was partly lost in the 1980s and 90s, when many media companies went public and turned into large financial structures, Cerezo said: “We need to recover the public service mission and offer a value proposition not only through content, but through the relationship with the audience.”

The goal is to deliver “a value proposition of services, products, and content that is as distinctive as possible,” capable of responding to the interests of the community, whether niche or multi-local.

Cerezo underscored the need for distribution and loyalty tools that enable a diversified monetisation model, always keeping the user at the center. He issued a warning: “If we don’t have an equation that responds to this ecosystem, it will be hard to survive in an increasingly complex environment.”

Walled gardens and the lesson of social media

Media companies were slow to grasp the concept of “walled gardens,” Cerezo said, while social media became highly effective platforms for extracting data and monetising user information. 

Meanwhile, news outlets lagged in registering users and building communities within those closed ecosystems.

He urged media companies to adopt the same logic as social media — and now AI — so they can build their own walled gardens, making it harder for external players to compete or operate within their ecosystems.

The end of the open Web? 

On a question of the open Web, Cerezo argued that media outlets joined this environment relying on “illusory traffic” driven by third-party platforms. This traffic was measured in terms of visits, not actual users or readers. 

As an example, he highlighted the functionality of Google Discover, which supposedly shows users personalised content but in reality prioritises “what the algorithm wants” without differentiating or building loyalty among audiences. 

“With the arrival of AI, if we don’t break this model, we’ll continue living in an open, free Web that will be irrelevant in monetisation terms,” he said. 

While saying he believes the open Web will continue to exist, he said outlets that depend on massive traffic without building communities or knowing their audiences will become increasingly irrelevant and struggle to monetise their products.

AI, a transformative agent

Cerezo argued that the arrival of AI is accelerating the shift toward a more closed ecosystem: “We’re at a stage where AI represents both a threat and an opportunity.” 

The industry has long lived with predictive AI (algorithms that anticipate browsing behaviour or tools such as Google Maps), he said. The major disruption now comes with generative AI, which lets users interact to create text, images, or audio and interpret data in a human-like way. 

While predictive AI transformed distribution and monetisation in relationships with third parties, Cerezo said generative AI alters the media’s DNA itself: the creation and aggregation of content. 

“If the value chain changes, the industry changes,” he said, warning that it’s nearly impossible to compete financially with the tech giants pushing these models. He called for a more disruptive approach from media companies. 

Although the issue is on media executives’ agendas, Cerezo criticised the lack of adequate strategy and planning. 

To address this challenge, he proposed three strategic levels:

• Industry-wide, he called on large media groups to develop a joint strategy for negotiating with AI platforms and defending the value of content as an industry rather than individually, to avoid letting AI remain purely extractive of media-generated data and user relationships.
• For smaller outlets, he urged alliances to help monetize the use of their content in training AI models.
• On a professional level, he encouraged adopting AI in daily tasks, but following clear industry guidelines.

Alliances and legal battles over data

Cerezo pointed out that there are already ongoing initiatives, such as partnerships and lawsuits by media companies against AI developers.

He mentioned deals between OpenAI (the creator of ChatGPT) and outlets such as Spain’s Prisa Media and France’s Le Monde, as well as The New York Times’ lawsuit against OpenAI over unauthorized use of copyrighted content. 

These partnerships must be designed with a medium- and long-term view, Cerezo said, avoiding exclusive dependency on major platforms, and even considering the development of proprietary, controlled AI versions for confidential uses.

Changes in how we search for information 

Another radical change he highlighted is how users now search for information online. Today, it’s often enough to type a question in Google and receive an AI-generated answer, reducing the need to visit multiple pages. 

This phenomenon, known as “zero-click,” means users no longer click on traditional search result links. An article currently ranking first on Google could see an 85% drop in clicks if its link is replaced with an AI-generated summary, Cerezo explained. 

Over the past year, this effect has already reduced traffic to outlets such as Business Insider, The Washington Post, The Wall Street Journal, and The Huffington Post, according to Similarweb data shown by Cerezo. 

He warned that media companies will have to get used to much lower traffic volumes — and therefore declining ad revenue. A study he cited estimated two years ago that the expected decline in advertising due to AI ranged from 18% (optimistic scenario) to 64% (pessimistic): “In a few years, this could be closer to the pessimistic scenario if we don’t act,” he cautioned.

He also noted that more and more clicks are being funneled to Google-owned properties instead of third-party sites, with many publishers depending on Google Discover for over 80% of their traffic. 

“I’ve never been a catastrophist, but I think we’re at a crucial moment of paradigm change that could seriously impact media companies that thought the open web would last forever.”

Proposals for survival

Among the alternatives to address this crisis, Cerezo proposed:

• Boosting subscriptions and membership models.
• Creating editorial products for specific audiences (newsletters, podcasts, etc.).
• Investing in community building, not just content.
• Strategically distancing from Google to gain editorial, technological, and commercial autonomy (not meaning that Google is the enemy).

He also highlighted the need to develop subscription models powered by AI, describing them as “the most natural relationship model for the survival of media.” These models should focus on user acquisition and retention, operational optimisation, and reducing churn. 

For audience segmentation, he suggested criteria such as device used, geographic location, level of engagement (new, frequent, passive subscribers), or content type (sports, politics, entertainment). 

He also proposed mechanisms such as offering a certain amount of free content behind paywalls, intelligent paywalls tailored to reader profiles, pricing and promotions based on geolocation and behaviour, and personalised visual design (buttons, forms, and messages). 

Smart paywalls adjust access to digital content based on user behaviour, location, reading history, or level of commitment, helping maximise conversions without hurting the user experience. Thanks to AI, subscriptions and advertising — previously seen as opposing worlds — are moving closer together to benefit publishers.

The human touch 

Cerezo addressed another critical topic: human involvement in content creation and automation processes, known as human in the loop. He emphasised that AI still makes mistakes, hallucinates, produces falsehoods, or plagiarizes, which requires human editorial oversight.

He also warned about the evolution toward agentic AI, which can make decisions autonomously with less human input. “This shift forces us to rethink the role of media, journalists, and employment in the future,” he said. 

Regarding advertising, Cerezo projected that by 2032, AI will influence 90% of all global advertising revenue, equivalent to US$1.3 trillion. These tools will enable testing of infinite ad variations to find the most effective one, thus eliminating human creativity in programmatic advertising. 

Cerezo called for ethics and governance in all these processes, urging that business models be reviewed by humans, “with rigor, truthfulness, transparency, and always putting the reader first.” 

He stressed the need for dedicated human teams to manage AI in media companies, even if the tools become fully autonomous, to keep outlets relevant with differentiated content and products.

. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .