The New York Times builds Blossom, a Slack bot that suggests which stories to post to social media; suggested posts typically get 380% more clicks
Shan Wang / Nieman Lab :
Context & Ripple Effects
Blossom is an early instance of the assistant-as-work-surface pattern: rather than shipping a reader-facing app, The New York Times embeds its recommendation engine where editors already work, inside Slack, and lets the bot rank which stories go out socially. The measured lift — suggested posts typically earning 380% more clicks — makes the case that curation-by-algorithm beats ad hoc editorial picks on distribution.
The surrounding coverage sketches the arc this sits in: months later Storyful was testing Slack integrations that push news alerts straight into clients' channels ([[a:836262]]), Facebook opened Messenger to chatbots in 2016, and by 2023 the idea had matured on both sides — BuzzFeed's Botatouille recipe bot turned the same suggestion mechanic into a consumer subscription play, while Slack itself productized channel summaries and search answers with Slack AI.
First-order effects
- NYT's social media editors shift from choosing posts themselves to triaging a bot-ranked queue, with Blossom's click-lift data deciding which stories reach Facebook and Twitter audiences.
Second-order effects
- Other newsrooms and intermediaries follow the workflow-embedded model — Storyful extends Slack from an internal tool to a client-facing alert channel, betting that delivery inside existing collaboration software beats separate dashboards.
- Platforms race to formalize the bot layer: Facebook launches Messenger's chatbot platform and releases Bot Engine built on its Wit.ai acquisition, turning bots from one-off internal hacks into a supported developer category.
Third-order effects
- Eight years on, the structure Blossom pointed at holds: Slack ships Slack AI natively, absorbing what early adopters hand-built, while publishers like BuzzFeed move the assistant from staff tooling to audience-facing products tied to subscriptions — the newsroom bot becoming both infrastructure and product.
The trend: News organizations are embedding assistants directly into collaboration surfaces, evolving from internal editorial helpers into platform-native features and consumer-facing products.