How to Make Your Website Visible to ChatGPT, Google AI, and Perplexity (What I Actually Did)

I talk a lot about building systems for other people. Today I turned that lens on myself.
I ran my own blog through two different AI-visibility auditing tools back to back, in the same afternoon. Not because I thought something was wrong. I was curious what an outside tool would say about a site I'd already spent weeks optimizing.
The results were humbling. Here's exactly what I found, what I fixed, and why I think every business — not just mine — should be doing this right now.

The Score That Started It

1st Tool gave my blog a 36 out of 100. "Poor."

Not because the content was bad. Content scored a respectable 60/100, and technical performance (speed, security) scored 90/100. The score-killer was one category: Website Data — 0 out of 100.

Zero.

"We checked 30 pages and found 0 data types that help AI understand your site."

That sentence is the whole point of this post.

What "Website Data" Actually Means

Here's the thing nobody tells you when you build a website: looking good to a human and being understandable to a machine are two completely different jobs.

A person reading my blog sees my name, my photo, my job title, my services. They understand instantly who I am and what I do. A crawler — the kind of system Google, ChatGPT, Perplexity, and every other AI tool uses to read the web — sees none of that automatically. It sees a wall of text and images with no labels.

Schema (technically "structured data") is the fix. It's invisible code sitting behind your visible content that explicitly tells machines: this is a person, this is their job title, this is a service, this is a price range, this is a question with this exact answer.

No schema means no labels. No labels means AI tools are guessing — or worse, ignoring you entirely in favor of a competitor who did label their content.

What I Actually Found (The Embarrassing Part)

My main site, srimarketer.com, already had solid schema — Person, Organization, Service listings, FAQ content. I'd built that weeks ago.

My blog subdomain had none of it. Not one line.

Two properties, same business, same person — and one of them was completely invisible to AI tools while the other was fully readable. I hadn't even realized subdomains needed separate treatment. Google and AI crawlers often treat blog.yoursite.com as a different entity than yoursite.com unless you explicitly tell them otherwise.

That one blind spot alone explains a meaningful chunk of the "0/100."

The Second Audit Surfaced Something Even More Specific

I also ran using my 2nd tool, a different tool focused on AI/generative-engine visibility specifically. It flagged a problem I genuinely hadn't considered: my name.
"Sri Krsna" is also a name associated with a Hindu deity — traditionally referenced as male. Every time I searched for myself on Perplexity, it referred to me as "he." Not because Perplexity was wrong on purpose — because nothing on my site explicitly told it otherwise.

The fix was structural, not cosmetic. Schema.org has an actual gender field built into its Person type. I added it. I also rewrote my About page in third person ("she," not "I") specifically so pronoun usage reinforced the correct signal in plain readable text, not just hidden code.

Two different problems. Same root cause: machines only know what you explicitly tell them. They don't infer, and they don't assume good faith.

What Fixing This Actually Looked Like

Over one focused session, here's what got addressed:

• Person, Organization, and WebSite schema added to the blog, matching what already existed on the main site — so both properties are recognized as the same entity, not two disconnected ones
• Service schema describing each specific offer, so AI tools can match search intent ("who audits business workflows") to an actual answer (me)
• Course schema for paid offerings
• FAQPage schema — eight real questions, real answers, covering everything from pricing model to data privacy to whether I serve clients outside my own country
• A gender field, closing the identity-ambiguity gap
• A dedicated, on-domain services page and About page — because it turns out even having genuinely good content on a different subdomain doesn't count if the crawler auditing this specific one can't see it

None of this changed what a human visitor sees. Every fix lives in the background. That's the part people miss — schema work is invisible, which is exactly why it gets skipped.

Why This Matters More Than Ever Right Now

A few years ago, this conversation was purely about Google rankings. It still matters for that. But increasingly, the audience reading your site isn't a human scrolling search results — it's an AI model deciding whether to cite you, recommend you, or quote you when someone asks it a question.
If your site has no structured data, you're not just ranking lower. You're functionally invisible to an entire, fast-growing category of discovery.
I didn't fully appreciate that until I saw my own score. Zero out of a hundred is not a "needs improvement" score. It's a "you don't exist here yet" score.

This Isn't a "Fix It Once" Job

Here's the part that surprised me most, even though in hindsight it's obvious: this isn't a one-time fix.

Anyone who's done real SEO work already knows this pattern. You don't optimize a page once and walk away — you monitor rankings, watch what competitors do, adjust when Google updates its algorithm, and keep checking whether what worked last month still works this month.

The exact same logic now applies to AI visibility — except the "search engine" isn't just Google anymore. It's ChatGPT, Perplexity, Gemini, Claude, and whatever comes next. Each one crawls differently, weighs signals differently, and updates on its own schedule.

So the plan going forward isn't "done." It's:

• Re-run the audits periodically (I'm checking back in roughly 10 days, then monthly after that)
• Watch whether AI tools actually start citing or recommending my site once the schema re-indexes
• Keep expanding structured data as I publish new content, not just patch the backlog once
• Treat this the same way I'd treat any Workflow Audit — as a system to maintain, not a task to close

If you fix your schema once and never look at it again, you're treating a living system like a checkbox. That's the mistake most businesses make with SEO in general, and it's about to become the same mistake with AI visibility.

If This Sounds Like Your Website Too

I'm not writing this to brag about a fix. I'm writing it because I genuinely think most businesses reading this have their own version of that "0" sitting quietly in the background, and most don't know it yet.

If you've read this far and you're wondering what your own site would actually score — that's exactly the kind of question a Workflow Audit is built to answer. It's the same systems-thinking approach I write about across this blog — see why fixing your workflow has to come before buying any AI tool for the bigger picture on that.

This post is part of Srimarketer Blog — honest lessons from building a business in the age of AI, written by Sri Krsna Sagun, Marketing Systems Consultant & AI Strategist. If you want to know where your own site actually stands, book a free 30-minute call — no pressure, just a real look at what's working and what isn't.