Inside the Shifting Realities of Modern Media: Key Insights from DGS25 

Every year, the Digital Growth Summit captures a moment in the media industry’s evolution. This year in Stuttgart, the moment felt particularly key, what with creators rising fast, AI redefining what it means to publish, and the decline of search.

The insights that emerged were simply too good to keep to ourselves. Here are eight takeaways that stood out. Follow Twipe Insights as we dive deeper into each one in the coming weeks 

1. Print’s Biggest Threat Isn’t What You Think 

In an era obsessed with being “digital-first,” print’s quiet endurance still surprises. On the CEO Panel, Felix Graf (CEO, NZZ) and Daniel Kempf (CEO, PD Digital) reaffirmed that print remains a core business and will stay highly profitable for years to come. Far from being a relic, print continues to anchor loyalty and revenue. 

But its survival may hinge on something far less romantic than readership: logistics. The real threat isn’t waning demand, but the collapse of physical distribution networks. As postal services consolidate and delivery systems shrink, getting the paper to the doorstep is becoming the industry’s weak point. The shift from print to digital, it turns out, might be decided not by audience behavior, but by the fragility of the supply chain. 

2. Your Biggest Competitor Isn’t Another Newspaper 

Nic Newman’s (Senior Research Associate, Reuters Institute of the Study of Journalism) presentation focused on how the news industry now fights on two fronts: automation from above, and creators from below. It’s the human creators — not AI — who are winning the attention war. Independent voices are building personal brands and direct trust with audiences, bypassing traditional media entirely. This trend varies by market. For instance, it is most advanced in the U.S., while legacy brands in markets like Germany still hold greater sway.  

Another challenge for publishers is the squeeze from above, with referral traffic from major platforms collapsing across all regions. According to the latest Reuters Insitute report, referrals are down 65% from Facebook and 50% from X over the past two years. This is yet another blow to the publisher’s advertising market and brings its own set of challenges to being discovered by audiences.  

3. AI is a Threat to Your Brand Disappearing

As strategist Nikita Roy (Founder & Host, Newsroom Robots Lab) explained: “The article is no longer the unit of journalism in an AI-mediated world.” AI models don’t read stories, but rather mine data. They erase tone, context, and brand, then repackage the facts in their own voice. In Roy’s words: “We’re looking at a future where your journalism is seen, but your brand is not.” In short, publishers risk becoming invisible data suppliers to the very systems that feed on their work and should invest in building brand loyalty on their own platforms.  

That need was echoed by Anne Tostain (Chief B2C Subscription Officer, Le Monde), who highlighted how even strong brands like Le Monde are grappling with the same challenge. Despite record search traffic in 2024 — with 50% of visits coming from search and only 30% directly — reliance on platforms remains high. In a workshop with attendees, she explored how product and marketing innovation can bring readers back home: from better curation, personalization, and live audio events to bold brand initiatives and stronger newsroom visibility.

4. The Hardest Part of AI Isn’t the Tech

While AI looms as an external threat, many speakers agreed that the toughest challenge is internal. Leaders like Alan Hunter (Co-Founder, HBM Advisory) and Gregor Zoller (Head of IT, MHS Digital) noted that resistance to AI adoption stems less from technical limits and more from cultural ones (e.g., fear of losing editorial tone, identity, and control). 

At MHS Digital, Zoller described how they tackled this by organizing internal “AI Days,” experimenting by launching sandbox tools like an internal “MHS-GPT,” and placing AI leadership roles at senior levels to ensure visibility. The key lesson is that transformation doesn’t start with technology. It starts with trust, literacy, and leadership that makes experimentation safe. 

5. Forget AI Assistants. The Real Future Is Autonomous Agent Systems That Don’t Need You. 

The most futuristic insight came from Markus Franz (CTO, Ippen Digital), who described a world where AI systems collaborate without human input. Instead of one assistant per task, imagine a network of specialized agents. For instance, an AI agent system could be composed of a “creative director” agent setting the vision, a “collector” agent sourcing assets, and an “editor” agent assembling the final story. 

This orchestration turns humans from operators (having an active role in the input) into observers (making sure everything runs smoothly). Franz demonstrated how Netflix already experiments with systems that generate audience-specific content entirely autonomously. It’s a glimpse into a near future where workflows are not just supported by AI, but run by it. 

6. How Publishers Are Already Building with AI 

Some publishers are already well past the experimentation phase. At Schibsted, Director of AI and Data Juan Carlos Lopez Calvet described a disciplined “Easy vs. Impact” framework guiding their development, choosing projects that are both feasible and transformative. That mindset has produced tangible results. Through a series of live demos, he showed how Aftonbladet’s evolving chat feature now guides readers through sponsored prompts, their “What Happened Since Last Time” tool re-engages audiences by summarizing updates since a user’s last visit, and VideoFly auto-generates short, multilingual videos from articles.  

At Frankfurter Allgemeine Zeitung (FAZ), Marina Sorg (Deputy CTO) shared how AI is woven into a broader strategy to strengthen loyalty and digital growth. Having reached 300,000 digital subscribers ahead of schedule, FAZ is now targeting 400,000 by 2030 — and AI plays a central role in that ambition. Three touchpoints already define their reader experience: an AI article summary that doubles engagement, an Eleven Labs–powered text-to-speech feature used by a third of subscribers, and an article chatbot being refined with close editorial oversight.  

7. What Europe’s Leading Newsrooms Agree On About AI

Despite their different markets, Le Parisien, The Telegraph, and Süddeutsche Zeitung shared a remarkably similar reality: the hard part of AI isn’t the technology, it’s the translation into newsroom value. Each publisher has moved beyond pilot projects into practical systems such as assistants that suggest headlines, chatbots that explain elections, automation that saves journalists time. Yet all emphasized that the real breakthrough comes from collaboration, not code. Whether through Le Parisien’s training sessions, The Telegraph’s hands-on prototyping, or Süddeutsche Zeitung’s co-creation model, trust and literacy are what turn AI from a tool into a teammate.

Another common thread was restraint. The panellists agreed that not every problem needs generative AI. Sometimes a simple metadata tweak outperforms a complex model. They also shared a cautious optimism: traffic disruptions from AI search features like Google’s Overviews haven’t yet proven catastrophic, but they’re being watched closely. Across all three organizations, the focus now is clear — use AI to deepen quality, reduce friction, and make journalism more human, not less.

8. The Real AI Challenge: Better Insights Don’t Mean Better Decisions 

In the closing keynote, Daniel Hulme, Chief AI Officer at WPP, argued that businesses don’t have insight problems, they have decision problems. True intelligence, he said, is “goal-directed adaptive behavior”: systems that learn from outcomes and improve over time. Most AI today isn’t intelligent but automated, and while current models think like “an intoxicated graduate,” they’re evolving fast: from master’s level reasoning today to “a professor in your pocket” by decade’s end.

Hulme urged leaders to focus less on flashy tools and more on intelligent decision-making. The value of AI lies not in faster answers but in better choices, but in building systems that explain themselves and optimize toward clear goals. As AI moves from automating tasks to augmenting judgment, the real competitive edge will come from how companies decide, not how much data they have.