Would an AI Even Quote You
Paste a paragraph — a blog intro, a product line, the last answer you wrote — and see whether an AI answer engine (ChatGPT, Perplexity, Google's AI Overviews) would quote you by name or quietly paraphrase you into its own words. The scorer reads your draft for four citation signals — a checkable figure, a crisp definition, short declarative sentences, and visible structure — turns them into a 0–100 quotability score, and names exactly which signals are missing. The moat is that last move: instead of "write better content," it points at the specific line a model can lift and the one thing you can add to make it liftable. Runs entirely in your browser (0 uploads, works offline).
How to use it
- Paste your paragraph into the draft box, or load one of the five seeded drafts (GST on food, SaaS retention, API latency, and more) to start from a polished-but-vague version.
- Watch the quotability score and the four-signal checklist update live as you type — each signal flips from ✕ to ✓ the moment your text earns it.
- Open the X-ray panel to read why a model would (or wouldn't) cite you, and the concrete fix for every missing signal.
- Hit Show a quotable rewrite to see the same idea carrying a checkable, structured claim — the version that scores 100.
What this clears up (the fundamentals)
- Classic SEO ranking and answer-engine citation are different games — ranking gets you onto a page of links; getting quoted gets your sentence lifted into the answer itself. The signals that win a citation are about the sentence, not the backlink.
- Checkable beats polished — a model paraphrases vague prose because there is nothing safe to quote. One specific, dated, checkable figure gives it a verbatim line it can attribute to you.
- Paraphrase without attribution is the default — when your claim is fuzzy, the model absorbs the idea and drops the source. Being uncredited is not bad luck; it is the predictable result of unquotable writing.
- Structure is a citation signal — a definition, a short declarative sentence, and a first/second ordering all map cleanly onto the shape of an answer, so the model can slot your line straight in.
Where it's used
Live on GEO/SEO and answer-engine-optimization resource pages, and as the L1 demo for the miatz module Positioning Backed by Retrieval. It is a miatz build-lab concept playable: play the scorer here to feel why one paragraph gets cited and another gets absorbed, then learn to build the citation rubric and the before/after reveal yourself.
FAQ
What is "answer engine optimization" (AEO / GEO)?
It is optimizing your writing so AI answer engines quote and cite it, rather than only ranking it in a list of blue links. Where classic SEO targets a search ranking, AEO/GEO targets being the sentence the model actually repeats — which is why checkable claims, definitions, and structure matter more than keywords.
Why would an AI paraphrase me instead of citing me?
Because vague, polished prose gives a model nothing safe to lift. It captures your general idea and rewrites it in its own voice — no attribution needed. A specific, checkable, well-structured line is quotable as written, so the model is far more likely to keep it intact and name the source.
How is the quotability score calculated?
Four signals, each worth 25 points: does the text carry a number or percentage (checkable figure), state plainly what something is (crisp definition), use short declarative sentences under ~25 words on average (direct sentences), and show order via a list or first/second wording (visible structure). Count the signals present, and that is your 0–100 score. It is a deterministic rubric — no model call, no randomness, no network.
Is anything uploaded?
No. Your draft is scored entirely in your browser — nothing is transmitted, stored, or logged. There is no live AI call: the scoring is a fixed on-device heuristic that runs with your Wi-Fi off.
What do the verdicts mean?
Skippable (under 50) — a model paraphrases you and moves on. Citable (50–99) — it might lift a line, but is still likely to absorb you. Quotable (100) — all four signals are present, and the paragraph reads like something a model quotes verbatim and attributes to you.
Limits
This is a teaching rubric with four fixed signals and precomputed, illustrative example drafts — it is a fast, opinionated read on quotability, not a live answer-engine crawl. It does not call a real model, so it cannot promise a given engine will cite you; retrieval, freshness, competition, and domain authority all matter too. The seeded figures in the example drafts are illustrative teaching numbers, not audited statistics. What the tool teaches, though — checkable claims and clear structure earn citations — is exactly the real practice.
Related
Part of the Demystify Playgrounds. Explore the rest from the Playgrounds home.
Bookmark this page (Ctrl+D, or ⌘D on Mac) — it works offline the next time you need it.