The Economist team has been explicit that a large part of the goal is understanding how journalism appears inside AI-native environments (like its new ChatGPT-native product) and reaching audiences where they increasingly spend time. ChatGPT now has hundreds of millions of active users, making it impossible for publishers to ignore as a discovery and information platform.  

This is now a marketing channel to show off their knowledge. There are several obvious advantages with this:

• Creates new distribution inside AI-native environments.

• Makes complex datasets easier to explore for consumers.

• Enables conversational access to journalism.

• Allows users to ask highly specific questions rather than navigate charts manually.

• Reinforces the value of proprietary publisher datasets.

• Provides learning opportunities around future audience behaviours.  

However, there are also limitations:

• Discovery remains dependent on ChatGPT rather than owned platforms (a persistent problem with channels that is unlikely to go away).

• Most publishers don’t possess unique datasets that are as well suited to conversational exploration.

• User behaviour around publisher apps inside AI platforms is still largely unknown.

• There is a risk these experiences become novelties rather than habits.

• The commercial model remains unclear.

• Publishers are still reliant on platforms they don’t control.  

What I find most interesting is what this signals about the next phase of publisher experimentation. For the last few years, publishers asked how AI would summarise articles. Increasingly the question has become: What if publishers build products specifically for AI environments? This is the first true example of this.

It’s possible, likely even, that there is a future where media organisations create small, highly focused experiences around their strongest assets — data, expertise, analysis, archives, or proprietary information — and distribute them directly inside conversational interfaces. 

The big question is around commercial models. Is something like this enough to drive consumer revenue? Or does it become a subscriber benefit inside a chat engine? The latter has obvious appeal but today remains out of reach.

What feels more significant is what this represents. For years, publishers worried about how AI would summarise their journalism. The Economist is asking a different question: What if journalism is designed specifically for AI environments?

This isn’t just a new distribution channel. It’s the first glimpse of a world where publishers build products for answer engines, not Web sites or apps. The winners may not be those with the most content but those with the most valuable expertise, data, and experiences to plug directly into AI-native platforms.

If you’d like to subscribe to my bi-weekly newsletter, INMA members can do so here.

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