A major American business daily is trialing an interactive assistant on its site that answers questions by pulling from its own editorial content. The feature, described by The Verge, integrates seamlessly with the publication’s library of articles to deliver responses grounded in its reporting and analysis.
Because the Ask FT database is built from Financial Times articles, the tool relies on that archive to surface the most relevant passages and context. For instance, when queried about the leader of a prominent artificial intelligence division, the system pointed to a recent appointment and framed the answer using FT’s reporting. The individual cited held the role starting just days earlier, on a specific date in 2024, illustrating how the bot ties answers to fresh editorial coverage while maintaining a clear source trail.
Responses from Ask FT appear with footnote markers in square brackets and include direct links to FT articles that underpin the bot’s statements. These citations provide readers with easy access to the original material that informed each answer, reinforcing the connection between the chatbot’s output and FT’s reporting standards.
At present, Ask FT is accessible to hundreds of FT Professional subscribers who rely on the service for business and finance insights. The chatbot runs on Claude, a large language model developed by Anthropic, and FT’s leadership has signaled ongoing experimentation with alternative models. The Verge notes that the underlying technology may evolve as the publication tests different language models to refine accuracy and usefulness for its audience.
There was a recent note about a built-in neural network used for photo editing on a well-known messaging platform, a reminder that AI integrations across media apps are expanding and evolving, sometimes sharing underlying technologies with those used for text-based assistants.