In our industry, tech moves quickly, but human nature stays fairly constant. Lately, we have noticed a trend among business owners looking to improve their search and AI visibility. They paste their website URL into a tool like ChatGPT or Claude, ask for an “SEO audit”, and end up with a multi-page list of automated fixes.

Some take it a step further, trying to find quick technical tricks to rank inside AI assistants – a concept people are calling Generative Engine Optimisation (GEO).

While AI tools have completely changed how we work, the foundations of search authority, business context, and brand trust remain human.

An AI audit isn’t useless, but it is incomplete. It looks at a website in a vacuum. It doesn’t know your business goals, your profit margins, or your competitors. More importantly, it can’t run a full strategic audit for you.

Can you perform a complete, professional-grade audit yourself using AI as an assistant rather than a driver? Yes, absolutely. But to do it properly, you have to plug the massive context and knowledge gaps that AI leaves behind.

Here is the exact process required to audit a website properly, along with the work involved if you decide to tackle it in-house.

Step 1: Manual Technical & On-Page Health Checks

While AI tools can scan website code for missing tags, they cannot evaluate how a website functions for a real person. A proper technical audit requires rolling up your sleeves and checking the site manually.

Shift To GEO And Digital Technology With AI Chatbots

What you need to do:

  • Test UX and UI across devices: Walk through the site on multiple screen sizes. Is the navigation intuitive? Do pop-ups block mobile buttons? Is the conversion or contact process smooth or frustrating? How does the UX/UI compare with competitors who are already ranking above you?
  • Verify indexation and rendering: Check how search engine bots crawl and render your pages. Look out for broken redirect chains, orphan pages, and slow-loading assets that waste crawl budget. Google Search Console can be used to cover 80% ish of this, to cover orphaned pages or complex redirect chains you need to use Screaming Frog or SEO PowerSuite for a full in-depth site scan.
  • Check for 404s and 301 redirects: Crawl the site to identify broken pages (404s) and review redirects (301s) to ensure they point to the correct, most relevant destination rather than a generic homepage fallback. There may also be redirects within a page that contain an incorrect internal link, which can be fixed by replacing it with the correct URL. Redirect chains and loops should be flagged and cleaned up, as they dilute link equity and slow down crawling.
  • Inspect content structure: Manually review page layouts to see if headings, internal links, and calls to action make sense for a human reader, not just a search bot. This includes:
    • Heading tag structure: Ensure that H1 to H6 tags follow a logical hierarchy rather than being used purely for styling.
    • H1 presence: Check that every page has one clear, unique H1 tag that reflects the page’s core topic.
    • Content depth on key pages: Assess whether important pages (service pages, category pages and landing pages) have sufficient, relevant content to satisfy both users and search engines, rather than being thin or duplicated across similar pages.
  • Review image ALT tags: Check that images have descriptive ALT text. This supports accessibility as well as image search visibility.
  • Audit meta tags: Confirm that title tags and meta descriptions are present, unique, and reasonably optimised for each page, rather than missing, duplicated, or auto-generated with no thought given to search intent.

What to study if you’re doing this yourself:

Step 2: Digging Into Real Traffic Data (GA4 & Google Search Console)

An AI prompt has zero access to your private analytics. It cannot see which pages actually drive revenue, where users drop off, or which search queries bring in real leads.

What you need to do:

  • Analyse Search Console performance: Look at impressions versus clicks, click-through rates (CTR), and average positions. Identify pages dropping in traffic or have within-striking-distance keywords (ranking position 8 to 20) that need a boost.
  • Check indexation status: Go into the “Pages” report (Index > Pages) in Search Console to see which URLs are indexed versus excluded. Key or important pages that aren’t indexed can’t drive any organic traffic at all, regardless of how well they’re optimised, so this is often the first place to look when traffic is lower than expected. Pay attention to reasons like “Crawled – currently not indexed” or “Discovered – currently not indexed”, as these often point to content quality or crawl budget issues.
  • Review your XML sitemap: Confirm the sitemap has been submitted correctly in Search Console, is free of errors, and only includes canonical, indexable URLs. An outdated or incomplete sitemap can mean important new pages take longer to be discovered and indexed.
  • Check for Coverage and Experience issues: Review the Page Indexing and Core Web Vitals / Mobile Usability reports for errors flagged by Google, such as redirect errors, server errors (5xx), or pages blocked by robots.txt that shouldn’t be.
  • Check for manual actions or security issues: Confirm there are no manual penalties or security issues (like malware or hacked content flags) sitting quietly in Search Console, as either of these can suppress rankings or traffic significantly without any obvious on-site symptoms.
  • Cross-reference Google Analytics 4 (GA4): Align search data with conversion data. A page getting 5,000 visits a month is useless if none of those visitors convert.
  • Identify keyword clusters: Group your search terms by user intent (informational, commercial, transactional) to make sure your pages answer what potential customers are actually looking for.

What to study if you’re doing this yourself:

  • Google Search Console has tons of different training series on YouTube – here is a good one https://www.youtube.com/watch?v=iB7_i4giItw
  • GA4 setup and reporting guides by Loves Data or MeasurementSchool on YouTube (you probably DON’T need the setup details, but the reporting guides are good – and the setup guides can help users understand GA4 more even if the process doesn’t need to be followed).

Step 3: Comprehensive Competitor Analysis

AI models cannot tell you why your top three local or global competitors are outranking you. They can only see what is currently on your page. To win in search, you have to reverse-engineer what is working for others in your exact market.

What you need to do:

  • Track rank positions: Use dedicated tools like SerpRobot or SEO PowerSuite to monitor exact daily rankings for your competitors across target locations.
  • Conduct a keyword gap analysis: Identify keywords your competitors rank for that your website does not currently target. Group these opportunities by search intent, relevance, ranking difficulty and potential commercial value.
  • Analyse competitor content and backlinks: Look at competitor domain authority, backlink quality, and page structures to spot content gaps and link opportunities.
  • Review technical setups: Compare site architecture, page speed, and schema markup against the market leaders in your niche.
  • Examine on-page SEO: Review competitor title tags, meta descriptions, headings, URLs, image optimisation, internal links and keyword placement. Pay particular attention to the pages consistently ranking above yours.

What to study if you’re doing this yourself:

  • SEO PowerSuite video tutorials on competitor research
  • Ahrefs or Semrush Academy courses on link building and competitive analysis

Step 4: Measuring Your Off-Site Digital Footprint (GEO & Surface Area)

If you want to appear in traditional search results and emerging AI recommendations (like ChatGPT or Perplexity), on-page tweaks are only half the battle. AI recommendation engines scan the web to see if real people trust, cite, and talk about your brand.

You do not optimise directly for AI by changing a few headings. You build your overall digital surface area.

What you need to do:

  • Audit brand mentions and reviews: Check your feedback on Yelp, Google Business Profile, Reddit, industry forums, and local groups.
  • Check NAP consistency: Verify your name, address, and phone number are 100% identical across your website, Google Business Profile and social platforms. Inconsistent details confuse both local search algorithms and AI engines trying to confirm who you are.
  • Verify citations and accreditations: Ensure your business details are accurate or are added to Chamber of Commerce sites, official trade bodies, and regional or industry specific directories.
  • Assess social integration: Evaluate how your active social media channels, video content, and external profiles feed into your overall brand presence.

What to study if you’re doing this yourself:

  • Local SEO guides by BrightLocal
  • Entity-based SEO and knowledge graph explainers on YouTube by experts like Jason Barnard

The Reality: Checklist vs. Strategy

True SEO And GEO Strategy

When you look at everything required – manual UI checks, GA4 data mining, keyword clustering, rank tracking, competitor profiling, and off-site audit work – it becomes clear why a single AI prompt falls short.

AI is a brilliant assistant. We use it every day to speed up data processing, spot trends, and expand on ideas. But using AI to generate a raw audit checklist without doing the manual strategic work is like asking someone to diagnose a complex car fault from a single photo. Occasionally it stumbles onto something useful, but most of the time it treats the wrong issue.

Taking on a full website audit yourself is entirely possible, but it takes days of focused effort, access to specialised software, and a deep understanding of how modern search works. It also means having an understanding on how to prioritise a checklist that tells you what is wrong, but not necessarily what to fix first – for the biggest impact on traffic and revenue. That prioritisation only comes from experience, and it’s often where self-run audits fall down, either fixing low-impact issues first or missing how different findings connect to one another.

We have some free tools linked from here https://teammarvel.co.uk/useful-links/ which can help with some of the above.

If you have the time to learn the tools and run the analysis, the steps above give you the roadmap to start to do it right. If you prefer to focus on running your business, that is where an experienced team with battle-tested processes could come in.