are the 5Ws and 1H dead, or the only thing saving technical writing from LLM fluff?Solved

Don’t panic just yet. You’re looking at the surface polish, not the actual substance. Sure, an LLM can generate a clean step-by-step list (the “How”). But it’s completely blind, deaf, and dumb when it comes to the actual context—and that’s where documentation actually succeeds or fails. 

The AI doesn’t inherently know Who your reader is. It treats a sleep-deprived DevOps engineer fixing a production outage at 3:00 AM (When) exactly the same as a non-technical product manager looking for a quick overview. It doesn’t understand Why a user should choose Option A over Option B; it just mindlessly lists both. If we let AI blindly dictate docs without anchoring them to human constraints, we’re just going to flood the tech world with “grammatically perfect” garbage that leaves users stranded. The 5Ws are the only thing keeping documentation user-centric right now. 

I think you both are hitting on the exact same problem from different angles. The 5Ws and 1H aren’t dead at all, they just got a massive promotion to your core prompt engineering framework.

Think about what happens when you give ChatGPT a lazy, generic prompt like “Write a guide on configuring X.” It gives you back massive walls of hallucination-prone fluff. But look at what happens if you feed it the exact context @jennifer just talked about:

• Who: Junior system admins.
• Why: Migrating legacy data securely.
• Where/When: Inside a restricted CLI environment during a tight maintenance window.

Boom. The LLM output instantly jumps from a generic 4/10 to a highly targeted 9/10. In this AI era, our value as technical writers isn’t in manually typing out the mechanical “How” anymore; it’s in mastering the context. If you don’t know your 5Ws, you literally can’t write a prompt good enough to survive in today’s tech space.