Macmillan staff say workplace AI tool NOA was rolled out then shut down

In short: Big publishers are testing AI tools at work, but some staff and authors say the tools are not helpful and are hurting trust.

What's going on

Macmillan Publishers staff told WIRED that in January 2024 the company introduced an AI chatbot called NOA, short for “No Ordinary Assistant.” Some employees said the rollout included flyers and a party at the company’s Manhattan office. Several assistants said they found the name offensive.

NOA was made by a German software company called Chaptr, which is backed by Holtzbrinck, Macmillan’s parent company, according to WIRED. Staff said they were told NOA was a “safe environment” for company information, meaning it was meant to be a closed system (like a private office chat, not a public social network). They were also warned it could “hallucinate,” which is when an AI tool makes things up, so humans had to double-check its work.

Employees said they were encouraged to use NOA for writing emails, drafting marketing text, and even creating goals for performance reviews. Several workers told WIRED the tool was unreliable and that “no one used it.” In October 2025, staff learned that NOA would be shut down. Workers said leaders later tried bringing Google Gemini into work processes, although Macmillan spokesperson Kelly Collins said there was no directive to use Gemini to evaluate agent submission letters.

What to watch

WIRED reports similar stories across the “Big Five” US publishers, including tensions with bestselling authors who want clear disclosure about when AI is used. The next key question is whether publishers set simple, consistent rules that protect authors’ work and keep humans responsible for final decisions.

Source: Wired