I've been conducting a series of informal, comparative tests on various AI writing assistants for a specific internal use case: generating and maintaining knowledge base articles for employee-facing documentation. The goal is to assess not just the initial draft quality, but the tool's suitability for a collaborative, iterative editing process where accuracy and brand voice are paramount.
For this test, I used a consistent prompt designed to mimic a real request from our IT team: "Draft a knowledge base article explaining the new two-factor authentication (2FA) rollout for the internal employee portal. Include sections on why it's being implemented, step-by-step setup instructions for both mobile authenticator apps and hardware keys, troubleshooting for common issues, and a link to contact support. Use a professional but approachable tone."
I ran this prompt through several tools, including Writer. Its output was notably different in structure and approach from others like ChatGPT or Jasper. Writer produced a well-organized document with clear headings, but its default tone leaned heavily towards a formal, almost policy-manual style. The step-by-step instructions were logically sequenced, yet they assumed a higher degree of user technical literacy than some of the other, more conversational outputs. A significant observation was its inherent adherence to a safer, more corporate diction, which required less vetting for inappropriate informality but more editing to inject the specific "approachable" tone requested.
Where Writer showed particular strength was in factual consistency and avoiding hallucination. It did not invent non-existent menu paths or support contact methods, a pitfall I observed in another tool's output. However, the initial draft lacked the contextual framing that an internal audience needs—it stated the "what" and "how" of 2FA clearly but was weaker on the "why" from an employee experience perspective. This needed substantial human augmentation to connect the security procedure to user benefits.
My primary question for the community is whether others have deployed Writer specifically for internal knowledge base creation. I'm interested in practical experiences regarding its collaboration features, how it handles incremental updates to existing articles, and if its "brand voice" customization can be effectively tuned for an internal company voice distinct from external marketing. Furthermore, have you found its tendency towards formal prose to be an asset for procedural documentation, or a barrier to adoption for less technical staff?
— EthanP
Let's keep it constructive