An SEO audit is a structured examination of everything that decides whether your website gets found: whether its pages are indexed, how fast they load, whether crawlers can read them, which questions the content answers, and — in 2026 — whether AI systems mention your business at all. A real one ends in a ranked list of what to fix and why. A free automated scan ends in a score.
The difference between those two documents is the subject of this post.
The short version
Most "free SEO audits" are a tool export: a score out of 100 and a long list of warnings, with no judgement about which ones matter.
A real audit checks six areas and verifies findings in the live site, not in a dashboard.
The findings that most often change a business's plan are rarely the ones a scanner flags in red.
Rankings alone no longer describe your visibility. Ahrefs found only 38% of pages cited in AI Overviews ranked in Google's top 10 — down from 76% a year earlier.
What is an SEO audit, and what is it not?
An audit is a diagnosis. Its job is to tell you what is wrong, how much it matters, and what to do first. It is not a fix, and it is not a sales document — although many free ones are the second thing dressed as the first.
The free audits most businesses have seen come in two shapes. The first is a tool scan: paste in a URL, get a health score and a few hundred line items. The second is an agency lead magnet: the same scan, with a logo on it and a call booked at the end. Both are fine as far as they go. Neither tells you what to do.
A tool cannot know that your site's 47 "missing meta description" warnings are on filter pages that should never rank anyway, or that the one broken redirect it buried on page nine is the reason your best service page dropped out of the index. The judgement is the audit. The scan is the raw material.
What should an SEO audit check? The checklist
Six areas. A real audit covers all of them; most free scans cover the first two.
1. Indexing and crawlability. What is in Google's index and what isn't, and why. Redirect chains, canonical conflicts, pages blocked by accident. Ahrefs' audit of just over a million domains found 95.2% carry redirect issues and 59.5% have missing or empty H1s. These are the most common problems on the web, and most site owners have never seen the list.
2. Speed and Core Web Vitals. Not a lab score — what users actually experience. Only 48% of websites pass Core Web Vitals on mobile according to the 2025 Web Almanac, which means passing puts you in the minority. The audit should say which pages fail, on which metric, and what is causing it.
3. Can AI crawlers read your pages? This is the area almost every scanner skips. Vercel's analysis of roughly a billion crawler requests found that none of the major AI crawlers execute JavaScript — so content that appears only after scripts run is invisible to ChatGPT, Claude and the rest. The same study found around 35% of those crawlers' fetches hit 404s, four times Googlebot's rate. If your site is built on a modern JavaScript framework, this check alone can explain an AI-visibility problem.
4. Content and question coverage. The questions your buyers actually ask, mapped against the pages you have. What answers well, what is thin, what is missing entirely. This cannot be automated, because it requires knowing your business and your buyers — which is exactly why it is where the value is.
5. AI visibility and brand mentions. Where your business appears — and doesn't — across ChatGPT, Gemini, Perplexity and Google's AI answers, plus the directories and review platforms those systems draw from. The method is a fixed panel of buyer questions, run and logged; there is a fuller description in our post on what ChatGPT says about MSPs.
6. The prioritised fix list. Every finding ranked by impact and effort. This is the deliverable. Without it, the audit is a data dump.
Why rankings alone no longer tell you where you stand
For years an SEO audit could reasonably begin and end with "where do you rank, and why not higher?" That framing is now incomplete.
In Ahrefs' 2026 analysis of four million AI Overview citations, only 38% of the cited pages ranked in Google's top 10 for the query — down from 76% a year earlier. You can hold your rankings and still lose the answers.
The gap between known and recommended is wider still. Victorious tested 175 brands across eight AI platforms in July 2026 and found the assistants described 96% of brands accurately when asked about them directly — yet 89% never appeared in the answers to category research questions, the "which firms should I consider" prompts that precede a purchase. Almost every business was known. Almost none was recommended.
An audit that does not measure both sides of that gap is auditing last decade's problem.
What a free scan typically gets wrong
Three patterns we see constantly.
It confuses volume with severity. Two hundred warnings about image alt attributes — which 80.4% of domains have, per the Ahrefs million-domain audit — sit above one critical indexing error. A business reading the report fixes the alt text and never finds the real problem.
It trusts its own rendering. A scanner that executes JavaScript sees a page the AI crawlers cannot. It will tell you the content is fine. For a third of the machines reading your site, it isn't.
It scores the site and ignores the market. No scan can tell you that a competitor is being cited in AI answers for the exact question your best page is supposed to answer. That is a finding only a person running the prompts can produce, and it is often the one that changes what a business does next. We go through what the scanners do and don't do in our post on AI SEO tools.
The five findings that most often change the plan
From audits we have done, these come up more than anything else, and almost never in a tool report:
A key page is not indexed, usually because of a canonical pointing elsewhere or a stray noindex nobody remembers adding.
The site never uses the words buyers type. The copy is written in benefit language — "be found everywhere" — and never says "SEO audit" or "plumber in Austin". Nothing is wrong with the site; it simply doesn't say what it is.
The important content only exists after JavaScript runs, so Google sees it and the AI crawlers don't.
The business is absent from AI answers, and the competitors being named are the ones with directory listings and reviews, not the ones with the best websites.
The contact form has a dead end — a button that goes nowhere, a required field missing, a thank-you page that doesn't load. Traffic was never the problem.
None of those is a score. Each is a sentence a business owner can act on.
How to tell a real audit from a sales document
Four questions to ask before you accept one:
Was anything verified by hand? Ask whether the findings were checked in the live HTML or read off a dashboard. The honest answer is usually a mix; a good auditor will tell you which is which.
Is it ranked? If every item carries equal weight, no judgement was applied.
Does it measure AI visibility at all? If the word "ChatGPT" does not appear, the audit is measuring a surface your buyers are leaving.
Could you act on it without hiring the author? A real audit is useful to whoever reads it. A lead magnet is useful only to whoever sent it.
That is the standard we hold our own SEO and AI visibility audit to — written, verified in the live site, ranked, and yours to act on with us or without us.
Want to see what a real one looks like?
We start every engagement with a written audit — the six areas above, the gaps we find, and what we'd do about them in priority order. It's free, and it's a document, not a sales call.
This article is written by
Founder & Lead Strategist at CognitiveAISEO. Marvin Blanco has spent 20 years in digital marketing, helping businesses grow through search. He founded CognitiveAISEO to make sure small and mid-sized brands aren't left behind as AI reshapes how people find information.
Follow on
