Associated Press now writes 3K stories per quarter on US earning reports after teaming with Automated Insights, plans to expand globally
Ross Miller / The Verge :
Context & Ripple Effects
In early 2015 the Associated Press is scaling what began as an experiment with Automated Insights' Wordsmith platform into core wire output: 3,000 automated US earnings-report stories per quarter, up from a few hundred written by hand, with a stated plan to take the same template global. Weeks later private-equity firm Vista Equity Partners moved on the startup behind it, acquiring Automated Insights outright (Vista's acquisition of Automated Insights), signaling that templated financial journalism had become a market rather than a novelty.
The arc since then has run one direction: France's Radar service reached thousands of automated stories a month by mid-2019 (Radar's expansion of automated story volume), and the AP itself turned its archive and automation experience into leverage for a two-year OpenAI deal in 2023 (the AP–OpenAI licensing agreement) — the same content pipeline this article first put on the map.
First-order effects
- AP business editors can redirect staff off routine quarterly recaps onto enterprise and analytical work, while coverage breadth jumps because every covered company gets a story regardless of size.
- Automated Insights gains a marquee reference customer at wire scale, making it immediately attractive to buyers like Vista, which acquired the startup within weeks.
Second-order effects
- Rival wires face pressure to automate their own earnings coverage or concede volume and speed on the most template-friendly beat in financial journalism.
- A vendor market forms around natural-language generation platforms, pulling in media companies beyond finance as Radar's monthly output demonstrated four years later.
Third-order effects
- Newsroom economics reprice toward editing and data infrastructure rather than drafting, and content archives become licensable assets — a path the AP followed through Civil and ultimately its OpenAI training-data deal.
- If automation normalizes for formulaic beats, disclosure standards and labeling rules for machine-written news become a regulatory question rather than a newsroom policy choice.
The trend: Wire services are converting formulaic reporting beats into software pipelines, turning newsrooms into editors of generated copy and their archives into licensed training data.