AI writing tools for authors

{{whyLabel}}: To maintain authenticity, you must first quantify what makes your writing sound like 'you' so you can guide the AI effectively.

{{howLabel}}:

• Analyze three of your best past writings.
• Identify your average sentence length (short/punchy vs. long/lyrical).
• List 5 'forbidden words' you never use and 5 'signature words' you love.
• Note your preferred perspective (e.g., intimate first-person vs. detached third-person).

{{doneWhenLabel}}: You have a written 'Style Guide' document summarizing your tone, rhythm, and vocabulary.

{{whyLabel}}: Not all AI models handle creative nuances equally; some are more 'robotic' than others.

{{howLabel}}:

• Use Claude 3.5 Sonnet for its superior ability to mimic complex prose and follow stylistic instructions.
• Alternatively, use GPT-4o for structural brainstorming and logic.
• Avoid smaller, older models which tend to rely heavily on clichés.

{{doneWhenLabel}}: You have an active account on a platform providing access to a state-of-the-art Large Language Model (LLM).

{{whyLabel}}: A system prompt acts as a permanent filter, ensuring the AI doesn't default to its generic, helpful-assistant tone.

{{howLabel}}:

• Draft a prompt starting with: 'You are a writing co-pilot for [Your Name]. Your task is to assist in drafting while strictly adhering to the following style...'
• Paste your Voice DNA into this prompt.
• Explicitly command the AI: 'Avoid flowery metaphors, do not use the word "tapestry" or "delve", and keep the tone [Your Tone].'

{{doneWhenLabel}}: You have a 200-500 word prompt that successfully generates text resembling your style in a test run.

{{whyLabel}}: Using AI for research saves hours of manual searching, allowing you to focus on the creative narrative.

{{howLabel}}:

• Use a tool like Perplexity AI or the 'Search' feature in ChatGPT to find historical facts, technical data, or setting details.
• Ask for 'counter-intuitive facts' or 'rarely discussed details' about your topic to add depth.
• Cross-reference critical facts using a traditional search engine to ensure accuracy.

{{doneWhenLabel}}: You have a structured research document with all key facts needed for your draft.

{{whyLabel}}: A strong outline prevents the AI from 'hallucinating' the plot or losing the logical thread of your argument.

{{howLabel}}:

• Start with a 3-act structure or a high-level chapter list.
• Use the AI to brainstorm 'what happens next' if you hit a wall, but make the final decision yourself.
• Break each chapter down into specific scenes or sub-points (approx. 3-5 per chapter).

{{doneWhenLabel}}: A complete chapter-by-chapter outline exists with brief summaries for each section.

{{whyLabel}}: Consistency is key to finishing; AI makes this goal achievable in under 60 minutes.

{{howLabel}}:

• Schedule a fixed 60-minute block each morning.
• Use the AI to 'expand' your rough bullet points into full paragraphs.
• Focus on quantity over perfection during this stage.

{{doneWhenLabel}}: You have completed your first 1,000-word session.

{{whyLabel}}: AI quality degrades over long outputs; drafting in small chunks keeps the prose sharp and focused.

{{howLabel}}:

• Feed the AI only the outline for one specific scene (300-500 words at a time).
• Provide 'Contextual Anchors': Tell the AI what the character is feeling or what the specific goal of this paragraph is.
• If the AI output is too generic, hit 'Regenerate' with a specific correction (e.g., 'Make this more cynical').

{{doneWhenLabel}}: You have a rough draft of your first chapter or major section.

{{whyLabel}}: This ensures the AI learns from your manual edits to improve future outputs.

{{howLabel}}:

• Take a paragraph you manually edited to perfection.
• Feed it back to the AI and ask: 'Analyze the style of this paragraph and describe it so you can replicate it in the next section.'
• Update your Persona System Prompt with these new insights.

{{doneWhenLabel}}: The AI's subsequent outputs require 20% less editing than the first ones.

{{whyLabel}}: AI has 'tells' (overused words) that signal to readers that the text was generated.

{{howLabel}}:

• Use 'Find' (Ctrl+F) to search for and replace/remove these common AI words: delve, tapestry, testament, vibrant, multifaceted, embark, intricate.
• Simplify overly complex 'purple prose' that the AI often generates to sound sophisticated.

{{doneWhenLabel}}: The draft is free of the top 10 most common AI-generated clichés.

{{whyLabel}}: AI cannot experience the world; only you can provide the specific, lived details that create authenticity.

{{howLabel}}:

• Identify 'flat' descriptions in the draft.
• Add one specific smell, sound, or unique personal memory to every second page.
• Replace generic metaphors with ones based on your own hobbies or background.

{{doneWhenLabel}}: Every chapter contains at least three unique details that an AI could never have known.

{{whyLabel}}: AI models 'hallucinate' facts that sound plausible but are entirely false.

{{howLabel}}:

• Highlight every date, name, statistic, or historical event mentioned by the AI.
• Verify each one using a reliable primary source (encyclopedias, official records).
• Correct any 'logical hallucinations' where the AI contradicted a previous chapter.

{{doneWhenLabel}}: Every factual claim in the manuscript is verified as 100% accurate.

{{whyLabel}}: Your ears catch rhythmic errors and 'robotic' phrasing that your eyes miss.

{{howLabel}}:

• Read the text out loud to yourself or use a high-quality text-to-speech tool.
• If you stumble over a sentence, it's likely an AI-generated structure that needs simplifying.
• Mark any sections that feel 'boring' or 'generic' for a final rewrite.

{{doneWhenLabel}}: You have listened to the entire draft and noted all rhythmic inconsistencies.

{{whyLabel}}: AI is an excellent developmental editor even if it's not a perfect prose writer.

{{howLabel}}:

• Upload your full draft to the AI.
• Ask: 'Identify the 3 weakest points in the character arc' or 'Where does the pacing drag?'
• Use this feedback to perform one final targeted edit.

{{doneWhenLabel}}: You have addressed at least three major structural improvements suggested by the AI.

{{whyLabel}}: Final errors can break the reader's immersion and trust.

{{howLabel}}:

• Use a grammar checker (like LanguageTool or Grammarly) for technical errors.
• Do a final manual pass specifically looking for 'voice consistency' from start to finish.

{{doneWhenLabel}}: A clean, error-free final manuscript is ready for export.