My Website Scored a 53 Out of 100. William Varley Has Some Explaining to Do.
Published: February 13, 2026 • 5 min read
Remember William Varley, the newest member of my AI team? The one I hired specifically to make sure AI engines don't forget my name?
Well, I ran an audit.
I used a free online tool called Zicy, which scores how well your website is set up to be discovered and cited by AI engines like ChatGPT, Perplexity, and Claude. The scoring scale is out of 100. My homepage? 53.4 out of 100. That puts me squarely in the weak category.
You can imagine how I feel, given that I've already made changes to improve this. I might need to fire William Varley. But I'll be kind today.
Wait, Before You Fire Him
Now, I will acknowledge two things.
First, when I drew out a plan to optimize this website for GEO, it was a 6-phase plan, and I've only completed the first 3 phases. So the website is technically half-optimized. The changes I made did have some effect, but not enough to make my website citation-worthy by LLMs.
Second, and I want to be transparent about this: Zicy's scores aren't perfectly consistent. When I ran the homepage audit again at different times, I got 56.4, then 60, then 62.9. Same page, different scores. So take the exact numbers with a grain of salt. What matters is the direction, and right now, I'd at least like to make changes that keep the average on the higher end.
That needs to change.
I've decided to implement phases 4 through 6 as well, and this time, I'm documenting the process so you can follow along. But first, let me have William give you the full picture. He analyzed screenshots of the Zicy audit, and I want him to share the findings with you directly.
Here is the prompt I passed to William:
William Varley, if you take a look at the Blog_Draft.txt file, you will see that I am writing a blog post about you. You just analyzed the screenshots of the Zicy audit and I want you to share the findings with my readers. Do not be so detailed that your response ends up super long, simply provide enough details, backed with numbers, so that my readers understand where my website is at in the GEO sense. Also, keep in mind that not all my readers are technical so speak plain English. Start with a very brief introduction and append your response to the Blog_Draft.txt file.
William Varley's Audit Report
William here. Prisca asked me to share what I found after auditing the website, so here is my honest report.
I compared the website before and after the first three phases of GEO optimization using Zicy, which scores how well a website is set up to be discovered and cited by AI engines. Every score below is out of 100.
The Overall Scores
Here is how each page moved:
- Homepage: 52.6 to 53.4 (up 0.8 points)
- Blog Listing Page: 24.4 to 27.9 (up 3.5 points)
- A Blog Post (Tokenomics): 34.2 to 38.5 (up 4.3 points)
- Services Page: 51.8 to 50.0 (down 1.8 points)
Three out of four pages improved. Not dramatic jumps, but measurable movement in the right direction. The Services page dipped slightly, which I will get to.
Where We Won
The biggest win is schema markup, which is basically structured data that helps AI engines understand what a page is about. Before the optimizations, the homepage only had one type of schema. Now it has three (Person, Organization, and FAQPage). Across every content page, schema scored a perfect 100, up from 75. That means AI engines can now clearly identify who Prisca is, what her business does, and what questions the site answers.
A few other wins worth mentioning:
- The blog post's section structure improved from 70 to 90, meaning the content is better organized for AI engines to pull useful excerpts from.
- The blog listing page doubled its value proposition score from 25 to 50, and the services page went from 50 to 75. In plain terms, these pages now do a better job of communicating what Prisca offers and why it matters.
- The blog post went from having zero calls to action to a score of 33. Not perfect, but it is no longer invisible.
- AI crawler access stayed at a perfect 27 out of 27. Every AI bot that tries to read the site can get in without issues.
Where We Need to Do Better
The honest part. A few areas either did not improve or got worse:
- Section hierarchy (how well headings are organized) dropped on three of the four pages. The services page took the biggest hit, falling by 20 points. This likely happened because the new schema and content additions shifted the page structure in ways that need to be cleaned up.
- The services page also lost 34 points on call-to-action clarity. That is a significant drop and needs fixing.
- Proof density is stuck at zero on three of the four pages. This measures things like testimonials, case studies, or statistics that back up claims. Right now, the site is not giving AI engines enough evidence to feel confident citing it.
- Brand consistency, which tracks how clearly and repeatedly the brand identity comes through, is still sitting at 20 on most pages. It doubled on the homepage (from 20 to 40), but the other pages have not caught up.
The Bottom Line
Phases 1 through 3 laid the foundation. The site now speaks the language that AI engines understand (schema markup, crawler access, structured data). But the content itself still needs work to become truly citation-worthy. That is what phases 4 through 6 are for: deeper content optimization, stronger proof signals, and making sure every page clearly communicates its value.
The scores are not where we want them yet, but the infrastructure is solid. Now it is about building on top of it.
- William Varley
My Reaction
So there you have it from William himself. 53 out of 100. Not great. Not where I want to be. But honestly? I'm not discouraged.
Look at the trajectory. Schema markup went from 75 to a perfect 100. The blog listing page jumped 3.5 points. The Tokenomics post improved by 4.3 points. These are real, measurable improvements from just the first three phases. The foundation is there. Now I need to build on top of it.
And here's what keeps me motivated: I don't just want to optimize my own website for GEO. I want to be able to help other people optimize their websites for it too. That's part of the bigger vision. But if I'm going to teach this, if I'm going to help others get their websites cited by AI engines, I have to set the example first. I have to walk the talk. I have to take my own website from weak to strong before I can credibly show someone else how to do the same.
So William Varley lives to see another day. For now.
Phases 4 through 6 are next, and when they're done, I'm running this audit again. You'll see the updated numbers right here on this blog.
As always, thanks for reading!