Companies introduce internal AI writing rules to curb low-quality output
Firms such as Polarsteps, Leapsome and Clay have issued formal policies that limit how employees use generative AI for internal documents, aiming to stop the spread of vague, overly long content.
Employers are moving from informal reminders to written policies that define acceptable internal use of generative AI. Polarsteps, based in Amsterdam, expects its 100-person team to label AI-assisted drafts and ensure the finished copy reflects personal reasoning. Leapsome’s mid-August policy for its Berlin and New York staff stresses brevity and personal accountability, while Clay’s August rollout requires engineers to stand behind every sentence they send.
The push follows a 2025 survey showing 40% of U.S. desk workers received AI-generated material that took almost two hours each to fix. Leaders say the rules have already reduced confusing output and are being positively received, with some staff urging even stricter limits on AI use in peer messages and marketing copy.
Why it matters
Clear AI guidelines help prevent wasted time and miscommunication in workplaces increasingly reliant on generative tools.
In this story
