Austen: the go-to-market guide

Simon, this is the whole commercial argument for Austen in one place: what you are selling, who buys it, at what price, at what margin, and what to do on Monday. It is written for you, not for a customer, so it says the uncomfortable parts out loud.

The whole thing in six lines

What. Austen turns a brief into a finished, on-brand article optimised for search and for AI answer engines, then a professional Proofed editor edits it and returns tracked changes.^S1
Who. Content agencies who resell, education and training providers, B2B SaaS teams hiring their first content person, professional services firms, publishers with commercial content desks, and e-commerce brands whose claims must be edited.^S38
How you get in. One SKU, one price, one page: 10 human-edited articles, $2,500, 30 days. It is built on the pilot branch, with a Stripe price command and a status page, and it is the only thing you lead with.^S3
What it becomes. Starter at $99 a month for 15 articles, Pro at $299 a month for 50, or a done-for-you retainer at $2,000 to $2,500 a month for eight finished articles.^S2
What it costs you. About $0.79 of model spend per finished article and about $53 of editing on a 1,500-word piece at Proofed's own published rate, so the pilot runs at roughly 79% gross margin.^S9^S10
Base plan. Month 6 at $10,988 from 2 pilots sold that month, 8 Starter and 4 Pro subscriptions live, and 2 retainers. Month 12 at $20,272.^S39

How this guide was made, and what it is not

Who wrote it. Simon Minton, with Claude doing the research legwork and the drafting, on 4 September 2026. The product description is read from the Austen repository itself, not from marketing copy. Competitor and marketplace prices come from each vendor's own pages, dated. Treat it as a well-researched briefing from a colleague, not as gospel.

What is solid. Everything about what Austen does today, because it was read from the code: the pipeline, the plan matrix in config/billing.php, the pilot SKU on the gtm/pilot-offer branch, the MCP server and its five tools, the model rate card in config/ai_pricing.php. Competitor prices and marketplace take rates, each with a URL and a check date.

What is judgement. The margin model, because the editing cost is an assumed transfer price rather than a measured one. The revenue scenarios, which are arithmetic on assumed win rates. Every funnel number in the sales kit. All of them carry a red pill on the sources page.

What this guide never does. It names no customer, quotes no testimonial, and gives you no usage number to use as a proof point, because Austen has no customer who has agreed to be named and no measured customer result. The local development database has zero users, zero articles and zero subscriptions, and there is no production access here.^S7 Where a proof point would normally go, this guide tells you to say "you would be early", which is a stronger line than a vague one.

What you already have that most people selling this do not

An editorial firm behind the copy

Proofed is a real editing business with a real bench and named clients on its own site: SoFi, Pinterest, Emeritus, Samsung, Substack, Forbes Books, Financial Times, Fujifilm, JP Morgan Chase.^S29 Every competitor in the AI writing category sells a draft. You sell a finished piece with a professional editor's name on the edit. That is the whole product argument.

A customer base you can email tomorrow

Proofed already sells managed editorial into eight published segments. Two of them, Marketing and Advertising and E-Learning, obviously need Austen.^S29 Those accounts already have a signed contract, a style guide on file and a named account manager. That is not a lead list, it is a book. Chapter 6 is how to work it without burning it.

A working product, not a deck

Austen is live at austenapp.com with Stripe checkout, a public REST API, an MCP server, agency accounts that join client workspaces without using a seat, WordPress publishing, one flow for improving an existing article and a demo-account reset. 751 commits between January and September 2026.^S1 Nothing in this guide needs building first except merging the pilot branch you already built.

The MCP server is a distribution channel, not a feature

Austen exposes an MCP server at austenapp.com/mcp with five tools: list-projects, draft-brief, generate-article, generation-status and get-article. Connecting is key-free: a client points at the URL, discovers the OAuth routes, registers itself and sends the user through an in-app consent screen. There is no token to copy and paste.^S6 It works today as a connector in Claude. Be exact about its reach: it drafts a brief and writes a scored article into the user's Austen account, which takes five to fifteen minutes, and ordering the human edit and publishing still happen in the app.

That matters commercially for one reason. The people most likely to buy an AI content engine in 2026 are already inside an AI assistant, asking it to write things. A connector listing puts Austen in front of them at the exact moment they ask, and it costs nothing to be there. Treat it as a channel with a task attached to it in the plan, not as a line on the feature list.

What is different about selling this

You are used to	In this market	So
Selling editing to someone who already has the document	The buyer has no document. They have a gap, a hiring freeze and a blog that has not moved in four months	Lead with the gap report, not the edit. The edit is what makes them trust the output once they want it
Pricing per word, transparently	The category prices per article and per month, and the cheapest credible competitor is $2.39 an article^S11	Never price per word to a marketing buyer. Price the outcome: ten finished articles, $2,500
A deadline that arrives with the file	You set the deadline. Nothing arrives unless you brief, chase and deliver it	The pilot is a 30-day delivery project, not a licence. Staff it as one
Buyers who know they need a proofreader	Buyers who have already tried ChatGPT and think they have solved it	The objection is not "does AI work". It is "why pay you". The answer is the editor and the brief, in that order (Chapter 9)
A market with a clear quality ceiling	A commodity software category with an $89 per article floor set by Verblio^S21	Never sell on price. You will lose. Sell on finished and on-brand
One buyer, one document	Agencies who resell, so one sale multiplies across their client list	Work the 25 agencies first. They are the only segment that buys repeatedly^S38
Reputation earned over years at Proofed	Austen has no reputation, no logo wall and no case study	Borrow Proofed's, precisely and honestly: name the editorial bench, never imply an Austen customer that does not exist

The chapters

What you are selling The product as built, including what it is not yet.
Market and competitors Every named competitor with a price and a URL, plus the ten marketplaces.
Pricing The pilot is the front door. Subscriptions, retainers, discount and invoicing rules.
Costs and margins Model cost per article from the rate card, editing cost, and the margin table.
The 12-month plan Three scenarios, a month-by-month timeline and what to do this week.
Warm paths Proofed's account managers and segments, and the rules for using them.
120 targets Sortable, filterable, one live trigger and a source per row.
Segment playbooks Six segments, nine parts each, from the buyer to the first 30 days.
Sales kit Qualification, discovery, demo, objections, a five-touch sequence, references.
Security, procurement and compliance Content ownership, plagiarism and AI disclosure, data handling, invoicing.
Metrics and the weekly review Twenty-four numbers and one hour every Friday.
Templates Pilot agreement, brief approval, acceptance criteria, renewal sequence.
Collateral Everything to send, with its HTML source next to the PDF.
Sources and assumptions Every number in the guide, with its type and check date.

How to use this

Read chapters 1, 3 and 9 before the first call

Product, pricing, sales kit. That is the minimum to hold a conversation without inventing something you will regret.

Ship the pilot SKU before the first email

The gtm/pilot-offer branch has the page, the Stripe price command and the webhook. It was cut from main on 4 September and main is now 71 commits ahead, so rebase and retest it, then merge it, run pilot:sync-price and set STRIPE_PILOT_PRICE_ID. Until that is done austenapp.com/pilot does not exist and every conversation ends in a bespoke quote.^S3

Work the agencies first, then education

The 25 agencies resell at their own margin and need no education about what a brief is. Education is where Proofed already has standing. Everything else is a new relationship.^S38

Overwrite the red numbers as soon as you have real ones

Every estimate in this guide has a red pill on the sources page and a line in Chapter 11 saying which measurement replaces it. After ten deals the guide should be mostly green.

The five things that decide whether this works

Whether the pilot can be bought without you. A page, a price and a button. Everything else is downstream of that.
Whether the human edit is visible. Tracked changes on the buyer's own draft is the difference between you and every $99-a-month tool. If the edit is invisible, you are selling Byword at six times the price.
Whether Proofed's account managers actually send the email. One sentence from a manager the account already trusts beats fifty cold emails. It is also the one asset you can only spend once badly.
Whether you can deliver 60 human-edited articles a month by month 12 without the quality slipping. The plan's base case gets there. Read the capacity note in Chapter 5 before you sell the fifth retainer.
Whether the credits meter tells you the truth before you sell volume. Measurement is on and enforcement is off, which is correct, but a done-for-you retainer can still lose money silently if nobody reads /admin/costs, or credits:usage once the pilot branch is merged, every Friday.^S7

What you are selling →