Search Experts Order an audit

SEO and GEO audit

We show what blocks indexing and visibility, why it matters and where to start the implementation.

Order an audit

In short

An SEO and GEO audit combines technical analysis, content analysis and an assessment of your brand’s visibility in AI-generated answers. The result is an action plan with an order of implementation.

What the service includes

Technical diagnosis

Crawl, indexing, rendering, architecture, performance and structured data.

Content and AI answers

We check whether your content answers customers’ questions and whether AI assistants name your brand and cite its pages as sources.

Implementation plan

A list of problems and tasks ordered by impact, urgency, cost and dependencies.

What we measure

The metrics stay the same from period to period, so every report can be compared with the previous one.

Problems blocking indexingin the report
Visibility gapsin the report
Citations and AI share of voicebaseline
Implementation priorities90-day plan

SEO and GEO audit: what it looks like and what comes out of it

Updated: Search Experts team

Contents
  1. What you get after the audit
  2. How an audit runs, step by step
  3. When to run an audit
  4. What we check in an audit
  5. How we set the order of implementation
  6. Problems we check first
  7. An expert audit and a tool report
  8. What an SEO and GEO audit is
  9. How to prepare for an audit
  10. What the scope and cost of an audit depend on

Order an audit and get a 90-day implementation plan

Order an audit

What you get after the audit

A report with the diagnosis and a 90-day implementation plan. Every task has a priority, a description of the problem, evidence, the expected effect on visibility and guidance for whoever will carry it out. We write technical tasks so that a developer can move them straight into their task system.

The report includes a baseline for further measurement: keywords in the top 10, domain visibility, indexing status, and your brand’s share and citations in AI answers. The same indicators then show whether the changes made a difference, also as part of regular measurement of your brand in AI.

How an audit runs, step by step

  1. Conversation and access

    Goal, market, competitors. Search Console and Analytics, read-only.

  2. Data collection

    Crawl, Search Console, performance, links, first measurement in AI.

  3. Analysis

    The cause of every problem and the evidence: a URL, a screenshot or a number.

  4. Priorities

    Impact, urgency, cost of the fix and dependencies.

  5. Walkthrough

    A meeting with the people who will implement the changes.

  6. Implementation check

    On request: we check that the changes work and that Google has picked them up.

How an audit runs. The result: a report with evidence and a 90-day implementation plan.
  1. Conversation and access. We agree the goal, the most important services or categories, the market and the competitors. We get read-only access to Search Console and Analytics.
  2. Data collection. A crawl of the site, Search Console data, performance measurements, the link profile and a first measurement of brand visibility in AI answers.
  3. Analysis. We combine data from different sources and look for the cause of every problem. Each finding is described with evidence: a URL, a screenshot or a number.
  4. Priorities. We order the problems by impact on visibility, urgency, cost of the fix and dependencies between them.
  5. Walkthrough. We go through the results in a meeting with the people who will implement the changes: marketing, developers or an agency.
  6. Implementation check. On request, we check that the changes were implemented correctly and that Google has picked them up. Further work on the recommendations can continue as ongoing work on your site’s visibility.

When to run an audit

  • Visibility or Google traffic is falling and nobody knows why.
  • You are planning a migration: a new domain, a new shop platform or a change of URL structure. The change itself is covered by site migration preparation.
  • The site is recent and does not appear for the search terms it should.
  • You are starting work with a new agency or specialist and want to know your baseline.
  • Competitors appear in ChatGPT or Gemini answers and your brand does not.
  • A year has passed since the last audit, or the site has gone through major changes.

What we check in an audit

The full checklist: 61 checkpoints in 10 areas. For each one we explain why we check it, and in the report every finding comes with evidence: a URL, a screenshot or a number. Not every checkpoint applies to every site. We agree which ones fit yours in the first conversation.

Indexing and crawler access

A page that Google does not add to its index will not appear in results, however good the content is. If the crawler cannot fetch a page, Google will not index its content.

Data sources: Search Console, a crawl of the whole site, server logs (if available)

  • Rules in robots.txt, line by line A single rule can block Google from crawling a whole category, or the files needed to display its content.
  • Noindex tags in the page code and in server headers Left over from a test version, they remove pages from results once Google crawls them again.
  • Canonical URLs on every page, and whether Google accepts them A wrong canonical URL can make Google show another page in results instead of the working one.
  • Every status in the Page indexing report, with sample URLs It shows which known pages Google has indexed, which it has not and why, so nobody has to guess.
  • Sitemap: does it contain only working, indexable, canonical pages A sitemap full of redirects and errors is a weaker hint to Google about which URLs matter.
  • URLs generated by filters, sorting, parameters and pagination Thousands of unnecessary URLs use up crawl time that should go to new products and articles.
  • Googlebot visits in server logs, section by section Logs show in detail which parts of the site the crawler actually visits and which it skips. Search Console gives crawl statistics in aggregate.
  • Server response codes: 404 and 5xx errors, soft 404s, response time A slow or unreliable server means Google fetches fewer pages.
  • Redirects: chains, loops, temporary redirects used instead of permanent ones Every unnecessary hop slows the crawler down, and a loop cuts a page off completely.

Rendering and JavaScript

A person sees the finished page in a browser. A crawler first receives the raw code, and some AI bots stop there.

Data sources: URL Inspection tool in Search Console, comparison of the source code with the rendered page

  • Whether content, headings and links are in the HTML without running scripts Some AI bots fetch only the HTML. Content added by a script is invisible to them.
  • Whether links are real anchors with an address A crawler follows only the addresses in links. It will not click a button driven by a script.
  • Lazy loading and infinite scroll on lists Products and posts further down the list may never reach the index.
  • Content in tabs, expandable sections and pop-ups Only what is in the page code counts. Content loaded after a click can be lost.
  • A screenshot of the page as Google sees it, for every template type It quickly shows whether the crawler sees an empty page or an error message.

Architecture and internal linking

Internal links tell Google which pages matter most to you. Without them even a good page stays in the shade.

Data sources: Site crawl, Search Console (Links report), analytics

  • Number of clicks from the home page to every important page A page buried deep gets less attention from crawlers and people.
  • Orphan pages that no internal link points to Without internal links they are harder for Google to find and to place within the site. Being in the sitemap does not guarantee they get crawled or indexed.
  • Link text: does it say where the link leads Descriptive link text helps explain the topic of the target page. “Click here” says nothing.
  • Menu, breadcrumbs and links in body text They spread the strength of the home page across the site.
  • Link distribution: do the pages you care about get the most Sometimes the terms and conditions or contact page gets the most links, while services and categories lag behind.
  • URLs: www and non-www, http and https, trailing slash, capital letters Several versions of the same address spread signals across duplicates.

Content and match with searches

Google and AI assistants show the page that answers a question best. We check whether your pages answer it at all.

Data sources: Search Console (queries and pages), keyword data, a specialist’s content review

  • Keyword map: which page should answer which search term Every important search term should have one clear target page.
  • Cannibalisation: several pages with a similar purpose shown for the same search term We check whether Google picks the right page and whether the overlap between pages lowers visibility.
  • Gaps: topics customers ask about that your site does not cover We check which important customer questions your site leaves unanswered and which sources search engines and AI assistants show for them today.
  • Weak and empty pages: empty categories, descriptions copied from the manufacturer, duplicates Google ranks individual pages but also uses site-wide signals. We check which weak pages to improve, merge or remove.
  • Page titles and descriptions in search results They decide whether someone clicks your result. We check for duplicates and truncated text.
  • Headings and text layout Clear headings help a customer find the answer quickly. We check whether the layout follows customers’ most important questions.
  • Freshness: prices, dates, service names, contact details Outdated information puts customers off and can end up in an AI answer as fact.
  • Pages losing impressions, compared across periods A drop noticed early lets you find its cause sooner.
  • Trust signals: who is behind the content, company details, contact, reviews Google’s guidelines for quality raters put heavy weight on this, especially for health and finance.

Performance and mobile version

A slow page tests the customer’s patience, and Google assesses a page mainly in its mobile version.

Data sources: Data from real Chrome user visits (Core Web Vitals), PageSpeed Insights, browser tests

  • Time to load the main content (LCP) for every template type A good result is 2.5 s or less for at least 75% of visits, on mobile and desktop separately. A sitewide average hides slow templates.
  • Response time to a click (INP) A good result is 200 ms or less for at least 75% of visits, on mobile and desktop separately. Delayed buttons and filters frustrate people while shopping.
  • Unexpected layout shifts on the page (CLS) A good result is 0.1 or less for at least 75% of visits, on mobile and desktop separately. A jumping layout causes accidental clicks.
  • What slows the page down: images, fonts, third-party scripts (chats, pixels, tags) Measurement shows which elements delay the page most, so it is clear where to start.
  • Server response time, caching, CDN A fast server speeds up every page at once.
  • Whether the mobile version has the same content, links and data as the desktop version Google indexes the mobile version. What is missing there is missing in Google.

Structured data

Schema.org markup describes the business, products and articles in a form machines understand, and gives a chance of rich results.

Data sources: Site crawl, enhancement reports in Search Console, Rich Results Test

  • Presence and validity of types: organisation, local business, product, article, breadcrumbs An error in product markup can remove the price and availability from results.
  • Whether the markup matches the content visible on the page Google requires structured data to describe what the user sees.
  • Errors and warnings in Search Console reports Critical errors show which items are not eligible for rich results. Warnings show what is worth improving.
  • Consistency of the company name, address and phone number with the Google profile and the contact page Conflicting data confuses search engines and AI assistants.
  • Expectations about markup Studies do not confirm that markup alone increases the number of AI citations. We say what it will and will not do.

Online shop, local business, multiple languages

Every type of site has its own pitfalls. We check the ones that apply to your site.

Data sources: Site crawl, Search Console, Merchant Center, Google Business Profile

  • Shop: which filters and product variants should be in the index and which should not Well chosen filters capture terms like “women’s trekking shoes size 39”. Badly chosen ones multiply duplicates.
  • Shop: unavailable and discontinued products A good product page has history and links. If there is a page with similar content, a redirect sends them there. Without a replacement, a 404 or 410 code is the right choice.
  • Shop: whether the product feed in Merchant Center matches the page Conflicting prices and availability limit product visibility in Google.
  • Local business: Google Business Profile (categories, opening hours, photos, reviews and replies) The profile decides presence in Google Maps and in local results.
  • Local business: location pages and how they connect to profiles Every branch should have a page that its profile points to.
  • Multiple languages: hreflang tags and language versions Errors send a customer from Germany to the Polish version, or the other way round.

External links and brand mentions

What others write about your company, and where they link to it, affects Google rankings and what AI assistants say about you.

Data sources: Search Console (Links report), link databases (Ahrefs, Majestic), mention search

  • Link profile: sources, topics, link text, sudden spikes It shows whether the strength of the domain rests on valuable sources.
  • Risky links and whether anything needs to be done about them Usually it is enough to ignore them. We act only when they do real harm.
  • Links pointing to pages that no longer exist If there is a suitable replacement page, a redirect sends those links to the right address.
  • Unlinked mentions, industry directories, comparison sites, reviews AI assistants build their picture of a company from sources like these.
  • Where competitors get links and you do not It gives a ready list of places worth appearing in.

Visibility in AI answers

People also use AI assistants such as ChatGPT or Gemini to look for information. We check what they hear there about your company.

Data sources: Our own measurement in ChatGPT, Gemini, Perplexity, Claude, plus AI Overviews and AI Mode in Google, Google Analytics 4

  • A set of customer questions built from real search terms We measure presence on the questions customers actually ask.
  • Whether the brand appears in the answer, in what context and next to which competitors This is the baseline against which we compare every following month.
  • Which pages the assistants cite as sources: yours or external ones It shows which content works and where you need to appear.
  • Whether the description of your company in answers is accurate and current A wrong price or location in an AI answer can mislead a customer. We also check where the assistant got that information.
  • AI bot access: robots.txt and firewall settings, for example in Cloudflare One setting that blocks AI bots can cut off Googlebot too.
  • Whether key information about the offer is written out plainly in text When the scope of a service, the area you cover and a “from” price are on the page, an assistant can take them from there. Otherwise it relies on outside sources.
  • Visits from AI assistants in Google Analytics 4 It shows which pages already receive people from AI answers.

Measurement and baseline

Without reliable data you cannot tell whether the changes made any difference.

Data sources: Search Console, Google Analytics 4, Tag Manager

  • Whether Search Console covers all versions of the domain and subdomains A missing property means missing data about part of the site.
  • Google Analytics 4 setup: events, internal traffic exclusions, filters Distorted data leads to bad decisions.
  • Baseline: keywords in the top 10, visibility, indexing status, brand share and citations in AI We use the same indicators afterwards to check the effect of each stage of the plan.

How we set the order of implementation

An order helps the team start with the changes that matter most for visibility. So we rate every task on four dimensions.

  • Impact: how many pages, search terms or AI answers the problem touches and how far it limits their visibility.
  • Urgency: whether the problem gets worse over time, for example when more pages keep dropping out of the index.
  • Cost: how much work, and whose, the fix needs: a change in a settings panel, a developer’s patch or a template rebuild.
  • Dependencies: whether other changes only make sense after this one.

Dependencies often reshuffle the whole order. If shop categories are out of the index because of a wrong canonical URL, improving their descriptions will not help until Google indexes them. That is why such a task goes to the start of the plan, even if it looks minor.

We divide the 90-day plan into stages. The first removes indexing blockers, the second sorts out the structure and content of the most important pages, and the third builds what affects the brand’s presence in AI answers: citable content, consistent company data and external sources.

Problems we check first

  • Important pages outside the index, or with a canonical URL pointing to another page.
  • Thousands of URLs from filters, parameters and pagination that scatter the crawler’s attention.
  • Several pages with a similar purpose, where we check whether the overlap between them lowers visibility.
  • Content that Google cannot read after rendering, or that AI bots fetching only the HTML cannot see.
  • Missing or wrong redirects after earlier changes to the site.
  • Service pages that do not answer customers’ basic questions, which makes them hard to quote.
  • Inconsistent company data in different places online, which confuses search engines and AI models.

An expert audit and a tool report

Free tools point out missing descriptions, titles that are too long or images without alt text. How much each one matters depends on the type of problem and the state of the site, and the tool does not know your site’s goals.

An expert audit uses the same tools as sources of data. The difference is in the judgement: we combine data from many sources, tell causes from symptoms and put tasks in order of their expected effect on visibility and the work they need.

What an SEO and GEO audit is

An SEO audit checks whether a search engine can find, read and understand your site without obstacles, and whether the content answers what people search for. A GEO audit extends that question to AI assistants: do ChatGPT, Gemini, Perplexity and Claude know your brand, name it in their answers and cite your pages as sources. This part is also covered by our AI visibility service.

We do both parts together, but we check visibility in Google and the brand’s presence in each assistant’s answers separately. For Google’s AI Overviews and AI Mode, a page must be indexed and eligible to be shown in Google Search with a snippet. Meeting these requirements does not guarantee that it will be shown.

The result of the audit is a plan: what to change, in what order and why. Every problem in the report has evidence, a description of its effect on visibility and the person or team who should fix it.

How to prepare for an audit

  • Read-only access to Google Search Console.
  • Access to Google Analytics 4, if it is set up.
  • A list of the most important services, categories or products and the main competitors.
  • Information about major changes in recent years: migrations, a platform change, a restructure.
  • A contact for the person who implements changes on the site.

What the scope and cost of an audit depend on

The size and complexity of the site matter most. A company site with a few dozen pages needs different work than a shop with thousands of products, filters and several language versions. Ongoing work on such a site is described on the e-commerce SEO page.

Other factors count too: the number of markets and languages, the platform and the way pages are rendered, the history of the site (migrations, penalties, drops) and the scope of the AI visibility analysis, meaning the number of questions and competitors to compare. We set the scope after the first conversation and a short review of the site.

Order an audit

Quote

We prepare the quote for your site, because it depends on its size and the scope of work. You get it after a short conversation, free and with no obligation.

How to ask for a quote

Fill in the form

Our experts have worked on SEO projects for:

  • Nestlé
  • DHL
  • Orange
  • VELUX
  • ERGO Hestia
  • ZnanyLekarz
  • Tubądzin
  • Grupa Pracuj
  • Toshiba
  • Vaillant
  • Enel-Med
  • ING Lease
  • Caparol
  • Tołpa
  • fru.pl
  • vola.ro
  • Archon
  • Moderna
  • NOYO
  • nice to fit you

Frequently asked questions

What do I get after the audit?

A report with the diagnosis and evidence, a task list ordered by priority, a description of the expected impact of each change and a 90-day implementation plan. We go through the results in a meeting.

Does the audit cover visibility in AI answers?

Yes. We check whether ChatGPT, Gemini, Perplexity and Claude name your brand in answers to your customers’ questions, how they describe it and which pages they cite as sources.

How is an audit different from a report from a free tool?

A tool detects issues and gives general guidance, but it does not know your site’s goals or situation. In the audit we assess which problems actually limit the visibility of your particular site, and we put them in the order in which to fix them.

What access do you need?

Read-only access to Google Search Console, and ideally to Google Analytics 4 as well. A list of your most important pages and the history of major changes to the site also help.

Does the audit require changes to my site while it runs?

No. The audit changes nothing on the site. We set the crawl so that it does not put load on your server.

Can you verify the fixes after the audit?

Yes. We can sign off the implementation, or carry on working together with the client’s team.

Does an audit make sense for a new site?

Yes, ideally before publication. In the test version we catch crawler access and indexing problems before the site reaches Google.

How often should an audit be repeated?

A full audit after major changes to the site, or once a year. Between audits, ongoing monitoring of visibility and indexing is enough.

Order an audit and get a 90-day implementation plan

In a short conversation we agree the scope: the size of the site, the markets and the questions to check in AI. After that you receive a quote.

Order an audit