Ask FT: Would you trust a news bot more if it knew less?

Ask FT is a useful Agentic Media case because it does not try to know the whole internet. It answers user questions by searching and summarising Financial Times journalism, then shows citations, links, and source extracts so readers can check the work. 1

The interesting design choice is the constraint: Ask FT is based only on FT journalism, searches the FT.com archive back to 2004, and is built for FT Professional subscribers rather than the open web. 2 FT Strategies says it became accessible to all FT Professional customers in April 2025, after internal launch and beta testing with a subset of users. 3

The carousel

Ask FT: a news archive starts behaving like a question-answering interface.
The move: the archive becomes the interface, not just a search box.
The design choice: FT Strategies says the team prioritised credibility and output quality over speed, with disclaimers and visible limits when the tool lacks enough information. 3
The agentic line: FT Strategies describes Ask FT as a RAG chatbot, not a full agent, but also treats source retrieval and answer generation as a foundation for more complex agentic tools. 4
The question: would you trust a news bot more if it knew less, but showed more receipts?

GIJN places Ask FT in a broader group of newsroom chatbots, alongside tools such as Rappler's Rai, Forbes' Adelaide, and Ask The Post AI, that answer from a newsroom's own reporting or vetted databases rather than general web search. 5

Discussion prompt: If a newsroom chatbot only answered from one outlet's archive, would that make you trust it more, less, or just differently?

Ask FT: Would you trust a news bot more if it knew less?

The carousel

참고 출처

댓글