The methodology, the measurements, and the mechanics — written for the founder who wants to know what they're actually buying, not just what the marketing says.
Search has split. A single score can't capture whether a business ranks in organic results, appears in AI-generated answers, and shows up when an AI assistant is asked a brand question. So we don't use one.
How well the site is set up to rank in traditional Google results. Technical health, structured data, entity clarity, on-page signals. The discipline is mature; the scoring is precise.
How likely the site's content is to be cited in Google's AI-generated summary results. When an AI Overview appears for a category query, the top organic result loses a significant share of its clicks — even when it still ranks first. This score measures the signals that influence whether you're cited in the AI answer.
How accurately AI assistants — ChatGPT, Claude, Perplexity and others — represent a business when asked about it. This surface is newer and harder to control, but the signals that shape it are knowable and improvable.
Each score is weighted equally. No score is more important than another — they measure different things. The overall is a single number you can track over time.
This number was set by auditing what exists in the live platform and verifying each measurement individually. We don't round up, and we don't count sub-measurements as top-level checks.
The 20 checks are discrete, named measurement areas — each one maps to a specific SEO, AEO, or GEO concern. Within each check, there are between one and six individual measurements (data points with their own pass/fail/score logic).
The 40+ figure refers to total individual measurements, not the number of checks. A check like "Schema markup" contains multiple measurements: presence, type coverage, required fields, JSON-LD validity, nesting depth. Each measurement is scored separately; the check score is the weighted result.
The AI brand interrogation feature is separate from the 40+ count. Interrogation is a distinct product feature — it runs separately from the audit pass and produces its own set of verdicts. Including it would misrepresent the audit score methodology.
AI brand interrogation runs independently from the audit. It asks five AI models real questions about your business — the kind a potential customer might ask — then grades each response for accuracy against what your business actually says about itself.
The five models are: ChatGPT, Claude, Perplexity, Gemini, and Mistral. Mistral is EU-based (Mistral AI, Paris) and was selected for GDPR-compliant data residency. These are the models that real buyers are already using to research businesses.
Each response is graded with one of four verdicts:
Where a model gets it wrong, the platform drafts a concrete fix — update press archive metadata, refresh your About page, add a structured ICP statement — and queues it for your approval.
What interrogation does not do: it does not claim to predict future AI citations or guarantee that fixing the issues will cause a specific model to change its response. It shows you what the models say today and gives you the levers to improve it.
| Plan | Interrogation included |
|---|---|
| Free audit | Not included |
| Starter | Monthly, 7 questions × 5 models |
| Growth | Monthly, expanded question set |
| Scale | Monthly, full question set + history |
Illustrative sample. Real interrogation runs monthly and uses your domain's actual content as the accuracy baseline.
Every other audit tool stops at the list. FixMySEO closes the loop: it drafts the fix, you approve it, and it deploys — then re-checks that the fix actually worked.
The audit runs across 20 checks and surfaces every failing measurement, ranked by expected impact on your scores.
For fixable issues, the platform generates the exact change needed — the new schema snippet, the corrected title tag, the structured FAQ block.
Nothing deploys without your explicit approval. You review the drafted change, edit if needed, and confirm before it goes anywhere.
The approved change is applied — via the WordPress plugin, snippet injection, or file edit, depending on your setup.
The specific measurement that failed is re-checked to confirm the fix took effect. Score updated.
The platform writes the fix and can deploy it directly once you approve. No manual implementation required.
The platform drafts the recommended change in full. You or your developer implements it — the draft removes the thinking work.
Some issues require decisions that only you can make. The platform flags them, explains the impact, and recommends a direction.
The scoring methodology behind FixMySEO is formalised in the AI Visibility Audit Standard — a published, open specification. This means you can audit our audit.
The Standard defines every check, every measurement, every scoring formula, and every weighting. It is published as an open specification and as an npm package — so any developer can read the schema, validate audit outputs against it, and verify that a FixMySEO score was calculated correctly.
The Standard is the reason we can say "40+ measurements across 20 checks" and mean it precisely. Every claim in the platform is anchored to a named measurement in the specification.
Read the Standard →The free audit runs your site through the full 20-check, 40+ measurement pass and returns three scores — SEO, AEO, GEO — plus a ranked list of fixes ordered by expected impact.
There is no account required. No email address. No OAuth. You paste a URL, the audit runs in under 60 seconds, and you get the results. No data is retained after the session unless you create an account and claim the result.
What the free audit does not include: the AI brand interrogation feature (that requires a paid plan and runs on a schedule), the auto-fix deployment loop (requires an account to save and deploy), and historical score tracking.
The free audit is designed to show you exactly what's broken. If nothing is broken, you don't need a paid plan. We think that's honest.
A platform you can trust is one that tells you where it stops. These are the things FixMySEO does not promise, and why we think being clear about this makes the platform more useful, not less.
FixMySEO fixes the structural and technical factors that influence visibility. Ranking is also affected by competition, domain authority, content quality, link profile, and algorithm changes — none of which the platform controls. The claim is: we fix what's fixable. Not: your ranking will go up by X.
The AI brand interrogation feature shows what models say about you and identifies the signals that improve accuracy. It cannot guarantee that ChatGPT, Claude, or any other model will cite you in a future response. The models make their own decisions. We improve the signals they read.
FixMySEO does the systematic technical and structural work that used to require an agency. It does not do content strategy, outreach, PR, or brand development. If you need those, you need those separately. For the 40+ addressable technical factors, FixMySEO handles them. For the rest, it gives you direction.
When a fix deploys, we re-check immediately to confirm it was applied correctly. Whether a search engine picks it up in two days or two weeks depends on crawl frequency and indexing — not on us. We confirm the fix is in place. We can't confirm when Google will notice.
The GEO score and AEO score both benefit from citations and quality inbound links. FixMySEO identifies where citation signals are weak and drafts the changes that make citations more likely to be accurate — but it does not conduct outreach campaigns or acquire links on your behalf.
The platform drafts structured fixes: FAQ blocks, schema snippets, corrected meta tags. It does not write blog posts, landing pages, or new site sections. The content strategy is yours — FixMySEO makes the content you already have work harder for visibility.
No account required. Paste your URL and get three scores and a ranked fix list in under 60 seconds.
← Back to homepage