The Pulse
Macmillan’s NOA Assistant Ended After Staff Called It Unhelpful
Macmillan introduced NOA, an internal AI assistant developed by Chaptr, to employees in January 2024 and Chaptr moved to sunset it in October 2025. Staff told WIRED that the tool was unreliable, while Macmillan says leadership did not direc
AI.info Team · Written with AI · fact-checked ·
Macmillan staff first met NOA in January 2024 through purple balloons, welcome banners and doughnuts arranged to spell out the chatbot’s name. The tool, whose initials stood for “No Ordinary Assistant,” was presented as a new member of the team at the publisher’s Lower Manhattan headquarters. By October 2025, staff learned in a company-wide email that developer Chaptr would sunset it. Employees told WIRED that few had found the assistant useful.
NOA arrived with a party and a warning
Macmillan introduced NOA as an AI chatbot made by German software company Chaptr, which is backed by Macmillan parent company Holtzbrinck, according to the email described by employees. The message characterized the tool as a safe place to enter proprietary information without sharing it with third parties. Staff were also warned in training that NOA could hallucinate and that its output needed human fact-checking.
Even with those cautions, employees say they were encouraged to use NOA to draft emails and copy, conduct marketing research and generate ideas. Several said they were told to enter only a “percentage” of authors’ manuscripts. Users could set the chatbot’s tone to options including “lighthearted,” “spartan,” “sarcastic” and “regretful.”
The rollout extended to performance reviews: later in 2024, an employee email promoted a specialized prompt to generate workers’ goals. Some human assistants, multiple employees told WIRED, found the name “No Ordinary Assistant” offensive. One staff member recalled a Chaptr representative emphasizing the importance of content to publishers, then joked, “And here I thought we were publishing literature.”
Employees say the tool did not earn its place
Several current and former Macmillan employees described NOA as unreliable and of little practical value. “It was terrible, and no one used it,” one employee told WIRED. The publisher’s planned sunset came after less than two years of staff-facing use, though WIRED’s reporting does not establish how widely employees tried the tool or how often they used it.
After NOA’s planned retirement, employees say Macmillan leadership began trying to integrate Google Gemini into publishing work. Some staff experimented with using Gemini to assess submission letters from literary agents; one editor described those evaluations as “useless.” Macmillan disputes that this reflected a leadership directive.
Macmillan rejects a directive to screen submissions with Gemini
Kelly Collins, Macmillan’s vice president of communications and community, told WIRED the company uses Gemini for its business but “there has not been any directive from Macmillan leadership to use Gemini for any part of the publishing process, including to evaluate submission letters from agents.” She also said the accounts “do not accurately reflect how NOA was rolled out and communicated,” without specifying which details she disputed. Collins said Macmillan’s approach to AI focuses on employee awareness and education while keeping people central.
The difference between employees’ accounts and the company’s statement leaves a practical question for publishing workers: whether staff are simply free to experiment with tools, or are expected to bring them into editorial and marketing tasks. The reports about Gemini concern experiments by some employees; Macmillan says leadership did not order its use to assess submissions.
Authors connect workplace AI to publisher trust
WIRED’s report places the Macmillan rollout amid broader unease among authors and staff at major publishers. Several bestselling authors interviewed by the magazine said AI use had strained their relationships with publishers; together, those authors have sold more than 30 million copies. One said marketing copy supplied by a publisher contained quotes that the author believed were hallucinated, while the publisher denied that AI had generated them.
Authors told WIRED they want publishers to disclose when and how they use AI, and one said they could not secure a promise from any Big Five publisher that AI would never be used on their books. Macmillan’s NOA experiment ended with Chaptr’s October 2025 sunset notice. Its later Gemini use remains disputed: employees described submission-letter experiments, while the company says it gave no directive to use the tool for that purpose.