Never ask an AI to “edit the book.”
Editing is a sequence of different questions. This week builds the sequence.
Your manuscript needs a dependable editorial floor.
Build staged manuscript workflows that produce genuinely useful reports, preserve your voice, surface continuity problems, and turn editorial decisions into an actionable revision queue — with you in charge of every meaningful change.
A live implementation week. Monday to Friday, 9–13 November 2026, taught by Elizabeth Ann West. Limited to 15 seats, because this is a build, not a broadcast.
The week promise
By Friday, you can run a manuscript through staged editorial reports, preserve the author’s voice, prioritise the work, and turn approved decisions into a revision queue.
Why “edit my manuscript” always disappoints
Handed one undefined instruction, a model does a little of everything and none of it well: it rewrites your voice while missing the continuity error, softens a line you fought for, and hands back something that is different rather than better.
The fix is not a longer prompt. It is a report-first sequence, where each stage asks one kind of question and produces one kind of artifact:
- Intake — version, size, project type, audience, goals, known concerns.
- Structure — promise, pacing, turning points, scene function.
- Story requirements — the author-defined elements that must develop, change, resolve, or stay consistent.
- Continuity — manuscript facts checked against series truth and the editing scope packet.
- Line and style — patterns, repetition, clarity, rhythm, voice risks.
- Copy and proof — mechanics, checked after substantive revision, not before.
- Revision queue — what to do, why, and where.
- Author decision — accept, reject, defer, or rewrite by hand.
Reports are not rewrites. The AI produces analysis; you produce the decision; the decision becomes a queue item. Nothing changes in your manuscript because a model thought it should.
The visibility rule that protects your voice
Editing reads the approved material, the relevant story context, and one immutable draft snapshot submitted from drafting. It does not browse alternate drafts, marketing, publishing, or your whole drafting folder. Reports stay in the editing zone; only a focused revision handoff goes back.
That single constraint is what stops an editing pass from quietly resurrecting a cut subplot or averaging your voice across four abandoned versions of chapter nine.
What you leave with
- A manuscript intake checklist.
- A staged editing workflow you can re-run on any book.
- A manuscript health report.
- A developmental report.
- An entity, continuity and project report.
- A line and style report.
- A prioritised revision queue.
- A human approval log.
- A reusable editing packet for the rest of the series.
The loop underneath everything
I keep my work in organised files, give the AI only the context needed for the job, ask it to produce a specific artifact, review the result as the author, and save the approved result in the correct work area. Only the approved current version is promoted into Final Files.
You do not need to remember a huge prompt. You need to know which file is authoritative, which supporting files are relevant to this job, what the AI is being asked to produce, how you will judge the result, and where the approved result is saved. That is the whole discipline, and it is why the system keeps working after the week ends.
How we talk about this
- The system is author-controlled, not fully autonomous. AI can organise, compare, analyse, suggest, transform and prepare. Approved truth and consequential changes pass through you.
- It is repeatable and maintainable, not effortless.
- It is software-independent — portable files, folders and workflow logic, not a proprietary platform.
- Approved information flows forward. Not every file is visible to every workflow.
- It scales capacity. It does not guarantee a specific income or publishing result.
- The books, genre, setting, characters, plot and audience stay entirely yours.
Or take the whole building
This week is one floor. The Ultimate Automated Author Experience is all five live weeks plus the self-paced classes and resources — bought separately it comes to $8,991.
The full experience is $4,999, with a preorder price of $3,499 until 12 September 2026. Limited to 30 seats.
THIS COURSE IS A PART OF THE ULTIMATE AUTOMATED AUTHOR EXPERIENCE
https://future-fiction-academy.teachable.com/p/the-ultimate-automated-author-experience-fall-2026
Who this is for
Working authors who are actively writing, revising, releasing or managing multiple projects; who want more capacity without losing creative control; who are willing to organise their files and make clear decisions; and who want repeatable workflows rather than one-off demonstrations.
You do not need to be a programmer. The system is built from folders, files, browser-based AI tools, subscription accounts and optional APIs. You do need a willingness to build the structure and maintain it.
Bring your own books. This week prescribes no genre, setting, characters, plot or subject matter. Bring a clean manuscript file, your series truth, and an editing scope packet.
Only working on one book? The system still works. The multi-book structure just shows you how to reuse the same environment when your catalogue grows.
Questions
When is it?
Monday to Friday, 9–13 November 2026. A live implementation week — demonstrations, Q&A and build time, not a pre-recorded course you watch alone.
How many seats?
Fifteen. You are building your own editing system with support, rather than watching someone else build theirs.
Will the AI rewrite my book?
No, and the system is deliberately built so it cannot. Every stage produces a report. You read it, you decide, and your decision becomes an item in a revision queue. Changes are applied on purpose, by you — never because a model preferred a different sentence.
Will this flatten my voice?
That is the failure mode the whole design exists to prevent. Line and style analysis identifies patterns, repetition and rhythm risks and hands them to you as findings. Nothing is smoothed into house style, and the editing workflow never sees the alternate drafts it could average your voice across.
Is this a replacement for a human editor?
No. It is a way to arrive at your editor — or your own revision pass — already knowing what is structurally wrong, what is inconsistent, and what order to fix things in. Editorial judgement stays human.
Do I need to be technical?
No. It is built from folders, files, browser-based AI tools, subscription accounts and optional APIs. You do not need to be a programmer.
Do I need specific software, or a specific AI provider?
No. Everything is portable files, folders and workflow logic rather than a proprietary platform. Use the tools you already prefer and replace them later without losing what you built.
What should I bring?
A clean manuscript file, your series truth, and an editing scope packet. This week prescribes no genre, setting, characters, plot or subject matter.
Do I need to have taken the earlier weeks?
No. Each floor has a standalone entry point. They are designed to compound, so if you want several, the full experience is the cheaper route.
