Answermend

Answermend / help

Help

Read this
before
you write in.

Six docs pages already answer most of what a check leaves you wondering, and this page routes to them rather than paraphrasing them — a summary that drifts from the page it summarises is worse than a link. Below that: the questions that aren’t about the measurement at all, and the one inbox for everything else.

Start here

The docs, in the order most people need them.

Each one covers a single question in full. Read the one that matches what confused you rather than all six.

What the check measures

The five steps and the four states, and why “named” and “read” are never added together into one number.

“Not measured” is not zero

What produces a dash instead of a result — a missing key, a rate limit, a blocked homepage — and why none of it is a verdict on your site.

Why two runs differ

Ask the same question twice and get two answers. Why one check is a sample, not a rate, and why there’s no percentage on a free result.

What to fix first

Five things, ordered by how certain each one is rather than by what it might return.

What this does not do

The surfaces nobody can measure, the numbers we won’t print, and what your scan data is used for.

All six, from the start

The index, with a short version of each page, if you’d rather browse than jump straight to one.

The first four of those pages are each about a screen, so here is the screen each one explains. All four captures come from one example scan — eight questions, four assistants, one run each, 32 checks attempted and 30 returned. The free check is one scan of up to 5 questions, one pass each, and only Gemini fetches on it: “cited” is measured on Gemini alone there, and on ChatGPT it is marked not measured — never printed as a zero.

app.answermend.com EXAMPLE DATA
Top of the report: 6 of 8 prompts named you, 4 of 8 cited you, 6 of 8 mentioned you, a dashed tile reading 2 checks not measured of 32, never counted as a miss, and a next-step card: 7 items follow from this scan, 3 to act on, with a What to fix button.
The top of the report, the screen What the check measures walks through: named on 6 of 8 prompts, cited on 4 of 8, mentioned on 6 of 8, and 2 of 32 checks not measured. What a description cannot carry: every counted tile prints the denominator it was counted against, in the tile itself, and “named” and “cited” are never added into one number. Northlight is an invented brand; its scan is simulated.
app.answermend.com EXAMPLE DATA
Four report tiles in one row: 6 of 8 prompts named you, 4 of 8 cited you, 6 of 8 mentioned you, and a dashed tile reading 2 checks not measured of 32 prompt-by-engine checks, excluded from all three denominators and never counted as a miss.
The dashed fourth tile is “Not measured” is not zero in one crop: the 2 checks of 32 that never ran are counted apart, excluded from all three denominators, and never scored as a miss. Northlight is an invented brand; its scan is simulated.
app.answermend.com EXAMPLE DATA
Question-by-question table for Northlight, one column each for ChatGPT, Gemini, Claude and Perplexity: cells read mentioned, cited, missing or not measured, each saying what happened on its one run, and a final column lists the domains named instead.
Why Why two runs differ exists: the same question gets a different verdict from each assistant — “Who are the best bookkeeping providers?” reads mentioned, cited, missing, cited across the row — and every chip carries the number of runs it was measured over instead of pretending one answer was the truth. A check that did not return is printed as not measured, never folded into a count. Northlight is an invented brand; its scan is simulated.
app.answermend.com EXAMPLE DATA
One What-to-fix item, chipped gap: questions answered without this brand in the answer. Do this: check that each question has a page answering it in its own words, where an AI crawler can reach it. Because: the brand was neither named nor cited in 2 of the 8 answered questions. Then the two questions, and a button to record a fix.
One item from the list What to fix first orders: what to do, why (neither named nor cited in 2 of the 8 answered questions), and the two exact questions it rests on, so the advice can be checked against its evidence rather than taken on trust. Northlight is an invented brand; its scan is simulated.

Not about the measurement

The other questions people arrive with.

Shorter than the docs pages, because these have shorter honest answers.

Is this free?

The tools are — all twenty, including the AI Visibility Checker, and they take no card. Two paid plans, Starter and Pro, are on sale; what each adds and what it costs lives only on the pricing page: nothing here is a price.

A check refused to run, or gave me a limit message

That’s a rate limit, not a result about your site. What’s enforced, why, and what a refusal actually says is on the acceptable use page.

What am I allowed to point this at?

Your own site, or one you’re authorised to check. The full boundary — what counts as misuse, and why the tools that fetch a URL are limited harder than the ones that don’t — is on the same acceptable use page.

Is there an account to sign in to?

Sign-in is open at app.answermend.com. An account is free and takes no card; the paid plans are bought by upgrading from the account page, through Paddle, and a subscription renews until you cancel it. Cancelling and refunds are one email — the how and the terms are on the refunds page.

Can other people in my company use this?

Yes, on every plan, and there is no per-seat charge. Invite them by email from the Team page in the app; each person signs in with their own link, so nobody shares a login. Roles are owner, admin and member — admins can invite and remove, and ownership can be handed to someone else, which matters when the person who signed up leaves. Removing somebody takes effect immediately.

Where does my scan data go?

It stays yours. It is never put on a public page, written up as a case study, or folded into any aggregate figure without your own separate, explicit opt-in — never bundled into a terms page. Full detail is on the privacy page.

None of this answered it — how do I reach a person?

One inbox, read and answered by hand. What’s worth writing in about, and what to put in the message, is on the contact page.

Still stuck?

One inbox, one person, replies written by hand rather than templated.

Write in