How I Used AI to Audit a Client's Website and Catch What the Developer Missed

1. The Client's Situation

My client is a professional who takes her brand seriously. She invested in a web developer to create a polished, functional website that matched her vision. She handed over her written copy, imagery direction, and layout preferences.

When the developer delivered the site, it looked decent on the surface. But my client had a gut feeling that things weren't quite right. She asked me to review it before she signed off.

That gut feeling turned out to be spot on.

2. Building the Crawler with AI

Rather than spending hours clicking through every page and inspecting elements manually, I used Claude (an AI assistant) to help me build a web crawler. The crawler was designed to visit every page on the site and check for specific quality markers.

Here's what the crawler looked for:

• ●Broken links and 404 errors across the entire site
• ●Missing or duplicate meta titles and descriptions
• ●Images without alt text (an accessibility and SEO issue)
• ●Pages with missing or incorrect headings hierarchy
• ●Placeholder text or lorem ipsum that was never replaced
• ●Inconsistent formatting and styling across pages
• ●Missing or incorrect contact information
• ●Pages that weren't mobile-responsive

The whole process took minutes, not hours. The crawler returned a structured report that I could walk through with my client line by line.

3. What the Developer Missed

The report was extensive. The developer had delivered a website that looked passable, but under the surface, it was full of issues. Some were cosmetic. Others were serious enough to hurt the client's credibility with potential customers.

Without going into specifics that would identify the client, the issues fell into a few categories:

• ●Copy that was provided by the client but never actually added to the pages
• ●SEO basics that were completely ignored. No meta descriptions, duplicate page titles, images with generic file names instead of alt text
• ●Broken internal links that led to dead pages
• ●Styling inconsistencies where fonts, spacing, and colours didn't match across sections
• ●Mobile layout issues where content was overlapping or cut off

This wasn't a case of subjective design preferences. These were objective errors that any professional developer should have caught.

4. The Client Got Her Money Back

With the crawler's report in hand, my client had clear, documented evidence that the work she paid for was incomplete. She was able to go back to the developer with a detailed breakdown of every issue. The evidence spoke for itself.

She got a refund for the project.

And she didn't have to guess or argue about whether the work met the standard. The data made it clear.

5. Fixing the Website Myself

Once the refund was settled, my client asked me to step in and fix everything. I went through the crawler's report systematically and addressed every issue:

• ●Added all the missing copy the client had originally provided
• ●Wrote and implemented proper meta titles and descriptions for every page
• ●Added descriptive alt text to every image
• ●Fixed the heading structure for proper SEO hierarchy
• ●Repaired all broken links
• ●Cleaned up styling inconsistencies across the site
• ●Ensured every page was fully responsive on mobile and tablet

The website went from a half-finished deliverable to a polished, professional site that actually represented my client's brand the way it should have from day one.

6. Why This Matters for Business Owners

Most business owners don't have the technical knowledge to evaluate whether a developer did a thorough job. They trust the professional they hired. But trust shouldn't mean accepting work blindly.

Having someone in your corner who can review the work objectively, using the right tools, makes all the difference. In this case, it saved my client money and protected her brand.

This is what a virtual assistant actually does at the senior level. It's not just admin work. It's protecting your business at every level.

7. The Power of Using AI as a Tool

I want to be transparent about something: I didn't write the crawler from scratch. I used AI to help me build it. And that's the point. AI isn't replacing the work I do. It's making me faster and more thorough.

The crawler I built in minutes would have taken hours to replicate manually. The value wasn't in the code itself. It was in knowing what to look for, how to interpret the results, and what to do about them.

Tools are only as good as the person using them. AI gave me speed. Experience gave me judgement.