Fold lifting one creased card from the glowing open drawer of a five-drawer cabinet

Visibility You can start this yourself today.

Answer engine optimization, measured and fixed.

The free audit is one output of a system that measures, diagnoses, fixes, watches and proves.

Every tool on this page serves one loop, from the question put to the engines to the fix you approve.

One mix-up, one fix

  1. Wire speaking into the microphone, then writing in its notepad

    The question

    Wire asks a buyer’s question.

    Wire puts a real buyer question to Perplexity and saves the reply as written, dated.

  2. Wire reading each card and dropping it into one of the trays

    The mix-up

    Wire spots the wrong brand.

    The answer was about a band, not the brand. Wire marks it and notes why.

  3. Murmur writing a line, reading the page, and laying it on the finished stack

    The fix

    Murmur writes the fix.

    Murmur writes the line that tells the engines which X this is, in the brand’s own voice.

Measure

One number for how the AI engines see you, worked out in the open.

The audit

The questions your buyers ask go to the engines, and the actual answers come back. Every mismatch with your verified brand facts is sorted by kind: a flat error is treated differently from a vague one.

Engine coverage

ChatGPT, Perplexity, Claude, Gemini, Grok and Google AI Overviews. If an engine doesn’t answer on a run, the report is flagged as degraded, not quietly smoothed over.

Brand IQ math

A rules-based reader scores how prominent you are in each answer, from zero to five: not mentioned, in the citation list only, in passing, a real option, the first sentence, or the first sentence called a top pick. Those blend into one number from zero to a hundred. Fixed, repeatable, with no AI judge.

Knowledge Check

Field by field (what your brand is, your mission, your founder story, your hero product, your audience, your values), did each engine get it right? You get a side-by-side grid of which engine knows you best.

Diagnose

From “the score is low” to exactly which claim, which engine, which question.

Gap cards

Wrong and misattributed findings are urgent: that is AI spreading a falsehood about you. They are ranked worst first by reach times damage: how many engines repeat the claim, times how harmful that kind of claim is.

Authority gap

For every question you’re missing from, Jinn reads who the engines cited instead of you and ranks those sites. Link building for AI answers, grounded in what the engines cite today, with “someone else was cited” kept apart from “nobody was cited at all”.

Prompt intelligence

Every tracked question is sorted into a buyer-intent family (comparison, alternatives, safety and trust, best-of, how-to, price and value), with an honest presence rate per family. An engine that returns nothing is not counted against you.

AI readiness

Five instant checks rolled into one verdict: your robots file, your llms.txt (a short file that tells AI crawlers what your site is), your sitemap, your structured data, and whether an AI crawler can reach your key pages at all. They run straight away, because AI crawlers don’t run JavaScript.

Fix

The audit ends in content, not a to-do list.

Corrections

Confirm a falsehood once and give the real answer. Every later audit, accuracy grade and drafted asset has to respect it, and the same claim is not flagged again. On purpose, a correction can’t move your Brand IQ.

Playbook

The gaps become a page-by-page rewrite plan: rewrite blocks for your existing pages, checked for voice and compliance before you see them.

Drafted assets

New FAQ, About and product copy grounded in your verified record: the words engines actually read when they answer. Plus the layer machines read: JSON-LD (structured data search engines use for rich results) and llms.txt.

Content calendar

Jinn reads the pages your tracked competitors rank for, groups the topics, finds where your coverage is thin, and turns the gaps into a schedule of what to publish next.

Fold, the Fox Uwee, folding feedback back into the brand record
Fold, Rules Editor (The Learner)Every piece of human feedback is folded back into the brand record, so it gets sharper with every campaign.

Watch

The score is a line, not a snapshot, once you turn the line on.

Monitoring

Weekly or daily re-audits per brand. Off by default: an honest empty state, not pretend always-on tracking.

Alerts and inbox

When a monitored run moves something that matters, such as a new falsehood or a score shift, it arrives as an alert, not a dashboard you have to remember to check.

Trends and tracking

Trends reads across your workspace; Tracking follows one brand. Both fill in after the first monitored run.

Weekly digest

The week’s movement in one email, once monitoring has run.

Traffic and revenue

Coming. A connector to Google Analytics and Search Console that ties movement in AI answers to your site traffic. It is built and switched off until Google verifies it.

Prove

Movement is measured, not asserted.

Proof

Re-run the audit and the lift is worked out run over run: the step that turns “we published the fix” into “the engines changed what they say”.

Board report

A frozen monthly report with the headline emailed on schedule: the version you forward, not export.

Brand Strategy in a Box

A document assembled word for word from your verified brand facts, with no AI writing it. A section you haven’t filled in doesn’t appear; nothing is invented to fill a page.

The evidence register

The discipline this page runs on: our published research is quoted and dated, the scoring formula is documented, and a pinned test proves the score ignores your own corrections.

Questions

What people ask about the features.

How often is my brand re-scored?

The free audit runs when you ask. With monitoring on, re-audits run weekly or daily per brand. Monitoring is off by default, and trends, alerts and the digest fill in after the first monitored run.

How does the corrections workflow work?

The audit ranks confirmed falsehoods worst first, by reach times damage. You confirm or dismiss each one. A confirmed correction is saved, respected by every later audit, grade and drafted asset, and the same claim is not flagged again.

Is Brand IQ a black box?

No. Prominence is a documented, fixed rules pass from zero to five per engine answer, blended into one number from zero to a hundred and rebalanced across whichever engines answered that run. The engine weighting is labeled a beta in the code and treated as a direction, not a precision instrument, and we say so.

Can I track my own questions, not only the defaults?

Yes. You can add your own tracked questions, and a discovery rail suggests real buyer questions, backed by search volume and grounded in your brand facts.

Can an AI agent run Visibility for me?

Yes. The Visibility agent server has 30+ tools on one connection, with tokens scoped to each workspace. An agent can list brands, pull findings, add tracked questions, start an audit and make a share link, and it gets a clear refusal at plan limits.

Bristle painting a jar, Ocho sorting folders, Wire on the phone and Proof checking a page, each in a lit arched window at night.

Visibility: questions about any part of it? Ask us.

  1. Tell us about your brand.
  2. A person on our team reads it.
  3. We read every application and reply within two business days.

Already use Claude or ChatGPT? Connect your AI, free