How do I write better AI prompts as an author?
- Context up front produces far better output than a bare request.
- Asking for critique beats asking for generation for most author tasks.
- Constraints — length, focus, format — sharpen the response.
- One question per prompt outperforms a stacked, multi-part ask.
- Iterating on a prompt is normal; the first version is rarely best.
Write better prompts by leading with context (genre, audience, what you are trying to do), asking the AI to diagnose or critique rather than generate, and constraining the output — a length, a single focus, a format. Ask one clear thing at a time instead of stacking requests, and iterate: refine the prompt based on what comes back. Specificity in equals usefulness out.
Chapter i·Why it matters
Most author disappointment with AI traces to the prompt, not the model: a vague "make this better" invites a generic rewrite, while a specific, well-framed request gets a targeted, useful answer. Because the same tool produces wildly different results depending on how it is asked, prompting is a learnable skill that determines whether AI helps your book or homogenizes it.
Chapter ii·What to include
- Context: genre, audience, and your goal.
- A critique-or-diagnose framing over "rewrite this".
- Constraints on length, focus, and format.
- One question per prompt.
- A voice anchor when the task touches your prose.
- Iteration — refine the prompt from the first result.
Chapter iii·Example
Instead of "improve this paragraph," an author prompts: "I write literary fiction for adult readers. Here is my voice sample. Tell me where this paragraph's rhythm goes flat — diagnose only, don't rewrite." The diagnosis is specific and usable, where the vague version would have returned a generic, voice-flattening rewrite.
WriteLoom structures AI around your book's context and voice, so prompts return targeted critique instead of generic rewrites.
See how WriteLoom uses AI