The audit takes the picture.Reading it correctly is the part you are actually buying.
SEO audit and assessment covering search and AI visibility in one pass. The audit gathers the evidence. The assessment tells you what it means.
Two different acts, and the market keeps selling them as one. Gathering the data is increasingly automatic. Interpreting it is where the exercise is won or lost, and where a wrong reading turns into months of work that moves nothing.
An audit and an assessment are not the same thing
The audit gathers
The state of the site as it actually is. What is there, what is missing, what the numbers say. A picture taken on a given day.
Much of this is mechanical, and more of it is automatable every year. It is not where the difficulty lives.
The assessment interprets
What the picture means. Which finding is the cause and which are symptoms of it, what is actually costing you money, and what is template noise that looks alarming and is not.
Recommendations come after this, and they inherit whatever it got wrong.
Where these engagements are won and lost
Why the interpretation is the whole exercise
A wrong assessment does not stay contained. It produces worse recommendations than no assessment at all. A misread cause becomes a confident plan, the plan becomes a quarter of work, and at the end nothing has moved and nobody can say why. The data gathering can have been flawless throughout.
It works like medicine, and the comparison is exact. Ordering the tests is not the skill. Reading them is. And reading them well comes from how many patients you have seen: the more cases you have worked through, the faster you recognise which symptoms matter and which are incidental.
The specific failure we guard against is naming the first explanation that fits. A set of data usually supports several readings, and the first coherent one feels like the answer, after which everything read afterwards quietly becomes supporting detail for it. So before we interpret anything, we write down at least two explanations consistent with the data and identify the exact field or report that would tell them apart. Then we go and get it.
That is a slower way to write an audit and a considerably cheaper way to be wrong.
One audit, six dimensions
Search and AI visibility in one pass, because the causes are the same
Being invisible in Google and being invisible in ChatGPT are usually the same problem wearing different clothes. Pages that cannot be crawled cannot be ranked or quoted. Product data that is incomplete fails both. An entity a machine cannot resolve is missing from both.
So we audit six dimensions in one pass: technical, on-page, off-page, local, answer engine and generative engine. Auditing them separately means paying twice to be told the same thing, and getting two recommendation lists that contradict each other on sequence.
Where they genuinely diverge, we say so. They are measured with different instruments and they do not move on the same timetable, and a report that pretends otherwise is not worth having.
Technical
Crawlability, redirects, canonicals, sitemaps, robots, speed and mobile. This comes first because everything below it is conditional on it. A page nobody can reach is not a content problem yet.
On-page
Titles, descriptions, headings, internal linking and duplication. Usually the fastest thing to move and the most commonly left exactly as it was at launch.
Off-page
Authority, backlinks, anchors and where competitors have links you do not. Reported honestly, which often means telling you this is not your bottleneck.
Local
Business profiles, name and address consistency, citations and local rankings. Marked not applicable when you have no physical presence, rather than padded out to fill a section.
Answer engine
Schema, structured questions and answers, and whether your pages are eligible to be used as an answer rather than just found.
Generative engine
Whether AI crawlers can reach you, whether you are citable, whether your brand is mentioned where models look, and whether your entity is clear enough to be resolved as one company.
What you actually receive
What we look at
Six dimensions, scored, with the evidence for each finding attached to it. Where we could not get authenticated access we say which platforms would have made the picture more complete, rather than quietly guessing.
Every methodology limit is stated in the document.
What you get
A score per dimension and a composite, the findings that matter with what each one costs you, and a sequence. Not four hundred issues. The ones that change something, in the order they unblock each other.
Written to be handed to whoever does the work.
What happens next
Usually nothing with us, and that is fine. If you want help it is normally a batch of work or a training rhythm rather than a retainer.
The audit is a complete piece of work, not a sales document with a score on it.
Scores are shown for every dimension, including the ones where you are doing fine. A report that only lists problems gives you no way to judge whether the problems are the important part.
White label audits for agency partners
For agencies: white label audits you can put your name on
We run audits for agency partners under their brand. You keep the client relationship, we do the diagnosis, and you deliver it as your own work.
This exists because it is a real gap. Plenty of good agencies have no in-house technical or AI visibility capability, and buying a tool export to hand over is not the same thing. The document format is built for it: a version with the fixes for you, and a client-facing version without them.
It suits you if you sell strategy, design or paid media and need the search and AI side covered properly, occasionally, without hiring for it. It does not suit you if you want a template report produced at volume, which is a different business and not one we are in.
What it will not do
What an audit from us will not do
- It will not give you four hundred issues. If everything is a priority, nothing is, and you already have a tool that can produce that list for free.
- It will not pretend every dimension matters equally for you. Local is marked not applicable if you have no locations. Off-page is reported as fine if it is fine.
- It will not hide how we know. Where we lacked authenticated access, the document says so and says what it would have added.
- It will not quote a number the data cannot support. Where the instrumentation itself is the problem, that is analytics and reporting rather than an audit finding. At low conversion volume most ratios are noise, and we will tell you which of your numbers are readable and which are not.
- It will not end in a retainer by default. Sometimes the finding is that you need one batch of work and then nothing for six months.
Who this is for, and who it is not
It fits businesses whose traffic moved and nobody can explain why, teams about to commission a large piece of work who want to know it is the right one, companies invisible in AI assistants, anyone inheriting a site or an account, and agencies who need the search and AI side covered under their own brand.
We will say no when:
- You want a cheap templated report at volume. That exists, it is a different business, and we are not in it.
- You already know the answer and want it endorsed. We will probably not endorse it, and you would be paying for an argument.
- Nobody will act on it. If there is no developer, no content owner and no budget, an audit becomes an expensive document. Better to wait.
- You want the fixes implemented on the same engagement. Diagnosis and delivery are separate deliberately, so the diagnosis is not bent toward the work that follows it.
Where this sits beside the rest
An audit is usually the first step into the rest of it: technical SEO when the foundations are the problem, on-page SEO when the pages rank but are not clicked, and ecommerce SEO when the catalog is the bottleneck.
For AI visibility specifically, see answer engine optimization and generative engine optimization. If you are spending on ads while you decide, the same diagnostic logic applies to paid search. The broader picture is on the SEO consultant page. And when the assessment concludes that nobody internally owns this, that is a fractional digital leadership conversation rather than a project.
Questions we get asked
Q. How much does an SEO audit cost?
It depends on the size of the site and how much of it you want covered, and we agree a fixed figure before anything starts. It is a bounded piece of work with a defined output, not a retainer and not a percentage of anything. The two things that move it most are the number of pages and whether you want the AI visibility dimensions included, and we will tell you on the call which parts are worth paying for on your site rather than selling you all six by default.
Q. Can ChatGPT do an SEO audit?
It can do parts of it well and other parts badly, and the difficulty is that the output looks equally confident either way. It is genuinely useful for reviewing content structure, drafting schema and explaining what a finding means. It cannot reliably crawl your site, it cannot see your Search Console or analytics data, and it will describe a JavaScript-rendered page from the HTML it was served, which is how confident and wrong conclusions get made about prices and stock. We use AI throughout this work. The judgement about what matters and in what order is the part that is not automatable yet.
Q. How do you do an SEO audit?
In dependency order, which is the part most audits get wrong. Crawlability and indexation first, because nothing below them matters until they are settled. Then on-page, then the structured data that decides whether you can be quoted, then off-page and local where relevant. Findings are gathered with evidence attached before anything is interpreted, and we look for a second explanation of every result before writing it down, because the first coherent story about a set of data is often the wrong one.
Q. What are the best SEO audit tools?
We use several, and the honest answer is that the tool is not the differentiator. Crawlers disagree with each other, tool severity ratings reflect what the tool can measure rather than what costs you money, and any two of them will give you two different issue counts for the same site. The value is in knowing which findings are real, which are template noise, and which one to fix first. If you want the tool output, most of them have a free tier and you do not need us for that.
Q. Do you audit AI visibility as well, or is that separate?
Same audit, same pass. Being invisible in Google and being invisible in an AI assistant usually come from the same causes: pages that cannot be crawled, product data that is incomplete, an entity a machine cannot resolve. Auditing them separately means paying twice to hear the same thing. Where they genuinely differ we say so, because they are measured with different instruments and they do not improve on the same timetable.
Q. Will you tell us things we do not want to hear?
That is most of the value. Common findings include that your paid budget is compensating for landing pages that should be fixed first, that your backlink profile is fine and not your problem, and that the work you were about to commission would not have moved anything. We would rather lose the follow-on work than write a report that leads somewhere convenient.
Q. Can you audit a competitor instead of us?
Yes, and it is a useful exercise, with one limit worth knowing in advance. Without authenticated access to their Search Console and analytics, a competitor audit is built from public data only, so it is stronger on technical, structural and AI visibility findings and weaker on anything requiring real performance numbers. We state that limit in the document rather than letting an estimate read like a measurement.
Q. Do you offer white label audits for agencies?
Yes. We run the diagnosis, you deliver it under your brand, and you keep the client relationship. The document set is built for it, with an internal version carrying the fixes and a client-facing version without them. It works well for agencies who sell strategy, design or paid media and need the search and AI side covered properly without hiring for it. It is not a volume report service.
Q. How long does it take?
Usually one to two weeks depending on the size of the site and how quickly access is arranged. The slowest part is almost always getting read access to Search Console and analytics, so we ask for it at the start. Without it the audit still runs, and it is weaker in exactly the places that matter most for deciding what to do next.
Q. What happens after the audit?
Often nothing with us, which is the intended outcome. You have a sequenced list your team can work through. If you do want help it usually looks like a batch of work with a control group so you can see what it did, or a training rhythm so your team can run it themselves. We would rather be the people you call twice a year than the people on a monthly invoice.
Start with a conversation
Thirty minutes on what changed, what you have already tried, and what you were about to spend money on. That last one is usually the most useful question.