Every growing business has someone whose judgment everyone wants on their work. At Every, a publication about AI that also builds software, that person is editor in chief Kate Lee.
The company roughly doubled to about 30 people in a year. One editor's time does not double with it.
What Every did
According to Platformer's interview with CEO Dan Shipper, Every collected a dataset of 30,000 of Kate Lee's historical edits, used it to build a copy editing agent and back-tested the agent against her past work.
Staff now tag the internal Every Agent and ask for a Kate copy edit on a piece or a landing page. Shipper says the agent keeps learning from what she corrects afterwards. He is candid about the limits: "It's not perfect, but it's much better than having her do everything." The interview reports no time or quality figures.
What it already gives the business
Every has turned one person's taste into something the team can reach without waiting for her. The editor stays in charge, because her corrections teach the agent what it missed.
It is a working example of expert knowledge captured from real work instead of a written manual.
The ExValu perspective, a hypothesis
Where a Company Brain could take it next
Editing is one skill. Picture the same approach applied to the reasons behind editorial decisions: why one headline worked, why a piece was held back, which readers responded to what.
A Company Brain could connect those notes with audience and product data, so a new writer understands the house view as well as the house style. That stays a hypothesis until it is tested against Every's own results.
What an owner and a buyer would look for
Track how long work waits for review, how often the expert corrects the same issue and how many reviews happen without them.
A buyer looks for exactly this: expertise that works for the team even when the expert is busy, plus a clear record of who approved the rules. Keep personal data of authors and customers out of training material unless there is a legal basis.
The problem behind this case
Facing the same problem?
About this analysis
* The image is AI-generated and does not show Every staff. The facts and figures on this page come from the published source named below.
This is an independent ExValu analysis of a published case. Facts, figures and names are as reported in the source below, which our team checked on October 12, 2026. The ideas for next steps are our own perspective and make no claim about conditions inside Every. To request a correction or removal, write to [email protected].
Primary source: Interview, independent journalism, Platformer, August 2026. Available at platformer.news/every-dan-shipper-interview-ai-writing.