AI SEO Audit — Find Out Why Google Can't See Your Website

Beautiful Website, Invisible to Google

I see this all the time — a client spends thousands on a gorgeous website with all the right information, but when their customers search for the business on Google, nothing comes up.

The problem isn't the design. It's SEO that was never set up in the first place.

An SEO audit used to require a specialist or tools like Ahrefs/SEMrush that cost hundreds a month. Now I do it myself with AI in 20-30 minutes.

The Tools I Actually Used

Primary tools:

• Claude Code — crawls the URL, reads source code, checks meta tags and heading structure
• Lighthouse CLI — measures Core Web Vitals and Page Speed (built into Chrome DevTools)
• Google Search Console — shows indexing status, coverage errors, and real-world Core Web Vitals from actual users

Alternatives if you don't use Claude Code:

• Screaming Frog SEO Spider — free for up to 500 URLs, crawls your site and exports CSV
• PageSpeed Insights (pagespeed.web.dev) — free performance measurement, nothing to install
• Ahrefs Free Tools (ahrefs.com/free-seo-tools) — checks broken links and meta tags for free

5 Things AI Checks For You

1. Are Your Meta Tags Complete?

Meta title and meta description are the first things Google reads. If they're missing or misconfigured — Google has no idea what the page is about.

What AI checks:

• Title tag — does it exist? Over 60 characters? Contains the target keyword?
• Meta description — does it exist? Over 160 characters? Compelling enough to click?
• Open Graph tags — so sharing on Facebook/messaging apps shows a proper preview
• H1 heading — only one per page? Aligned with the title?

The prompt I actually used:

Audit the SEO of website [URL] and report findings on:

1. Meta tags: title, description, Open Graph — complete? Too long? Contains primary keyword?
2. Heading structure: How many H1s? Duplicates? H2/H3 hierarchy correct?
3. Mobile responsive: font size, touch targets, mobile layout
4. Page speed: issues found, specify which elements are causing slowness
5. Keyword density: what keyword does this page seem to target? Where is it distributed?
6. Broken links: 404s or incorrect redirects

Format results as a table: Issue | Priority (High/Medium/Low) | Fix

2. Mobile Responsive

Google uses Mobile-First Indexing — meaning it evaluates your site from the mobile version first. If your site looks broken on phones, your ranking drops.

AI checks: whether fonts are readable on mobile, buttons are tappable, layout doesn't break.

3. Page Speed

Slow site = people leave = Google drops your ranking. Google requires passing 3 Core Web Vitals:

• LCP (Largest Contentful Paint) — should be under 2.5 seconds
• CLS (Cumulative Layout Shift) — layout shouldn't jump around during load
• INP (Interaction to Next Paint) — clicks should feel responsive

Measuring Page Speed with Lighthouse CLI:

# Install (if you haven't already)
npm install -g lighthouse
 
# Run an audit and export as an HTML report
lighthouse https://yourwebsite.com --output html --output-path ./seo-report.html
 
# Check just the performance score
lighthouse https://yourwebsite.com --only-categories=performance --output json | jq '.categories.performance.score'

AI tells you which metrics failed and what's causing it — uncompressed images, slow-loading JavaScript, etc.

4. Are Keywords Matching Your Target?

A site might have great content but use the wrong keywords. For example, customers search for "BBQ restaurant downtown" but the website only says "quality dining experience" — with no mention of location.

AI analyzes: keyword density, keyword distribution across headings and body text, internal linking between pages.

5. Broken Links

Broken links = bad user experience + Google doesn't like it. A 404 error on an important page will tank your rankings.

AI scans the entire site and reports every broken link, along with which page it's on.

What AI Delivers

The output isn't just a list of problems — it's an actionable fix list you can work through immediately:

The Actual Step-by-Step: Which Tool > Does What > Gets What

Step 1: Run Lighthouse to get a performance score and Core Web Vitals baseline

• Tool: Lighthouse CLI or PageSpeed Insights
• Input: the website URL
• Output: score 0-100 + list of issues with explanations

Step 2: Have AI check meta tags and structure using the prompt above

• Tool: Claude Code
• Input: URL + target keywords
• Output: issues table + fix list sorted by priority

Step 3: Have AI prioritize the fixes so you start with the highest-impact items

From the SEO audit results above, rank the fix list by impact (high -> low),
considering:
- Greatest effect on Google ranking
- Quick wins that don't require major code changes
- Direct impact on user experience

For each fix, estimate: time needed and which file to edit

Step 4: Have AI write new meta descriptions for every page that's broken

Write a new meta description for [page name] on a [business type] website
Primary keyword: [keyword]
Business USP: [differentiator]
Length: no more than 155 characters
Must be click-worthy + include the primary keyword in the first sentence

It Takes 20 Minutes

All you need: a website URL + the keywords you want to rank for.

AI will:

1. Crawl the site automatically
2. Check every point listed above
3. Produce a report with priorities and a fix list
4. If the brief is good enough, it'll even write new meta descriptions for you

The Bottom Line

SEO audits don't require a specialist anymore. AI covers all 5 areas — fast audit, clear fix list, act on it immediately.

If you have a website but traffic is low — start with an audit. Stop guessing where the problem is.

Further Reading

• AI Cost Analysis for Restaurant Menus — Know Your Real Margin Before You Sell
• Build a Startup Pitch Deck with AI