How I Built AI Search Visibility Over Time | An Agent-Ready Creator Case Study

I did not start changing the site because Shopify told me AI agents were coming

I started because I could already see the behaviour changing.

AI systems were sending measurable visits to JackRighteous.com. They were recommending individual resources. I found evidence of AI-assisted discovery reaching the purchase layer. At the same time, my raw traffic totals became less useful because bot-heavy and unattributed traffic could make the site look much larger without telling me whether real people were arriving with useful intent.

That forced a change in what I measured.

I stopped asking only, “How much traffic did I get?”

I started asking, “What did the visitor or AI system understand well enough to come here for, and what happened next?”

The traffic numbers were useful, but they were never the lesson

I have analytics. I watch referral sources. I compare search traffic, AI-assistant traffic, email traffic, internal clicks and purchase behaviour. But I do not want this case study to become a scoreboard.

The useful lesson is much simpler: traffic changes slowly, attribution is imperfect, and no single number tells you whether your site is getting better.

Over time I began seeing identifiable visits from traditional search engines and from AI assistants such as ChatGPT, Gemini, Claude and Perplexity. That was enough to prove that AI-assisted discovery was becoming a real acquisition channel for my site. It was not enough to prove that every visit was valuable, that every recommendation converted, or that one optimization caused the change.

SEO takes longer than most creator advice makes it sound

One of the easiest mistakes is to make three changes on Monday and expect Google, ChatGPT or another discovery system to reward you by Friday.

That is not how I have experienced it.

SEO is cumulative. A useful article needs time to be crawled, understood, linked, revisited and compared with competing pages. Internal links need time to create context. Search intent changes. Competitors improve. Platforms change how they surface information. Older pages can rise, flatten out or decline. A redesign can improve one part of the journey while accidentally weakening another.

That is why the work on JackRighteous.com has been iterative rather than a one-time “SEO project.”

There was no single trick that created the improvement

The changes came from many small practices working together. Some were traditional SEO. Some were content design. Some were conversion work. Some were simply removing confusion.

• Answering the main question sooner instead of making readers dig through a long introduction.
• Using clearer titles and headings so the page says what problem it solves.
• Building topic hubs so individual articles have a visible place in a larger system.
• Adding forward and reverse internal links so readers can move in both directions through a topic.
• Updating older useful pages instead of constantly creating new pages that compete with them.
• Improving product and offer language so a new visitor can understand the difference between free help, one-time access and consultation support.
• Watching engagement and conversion signals rather than celebrating traffic that does nothing.
• Testing AI visibility without abandoning Google, Bing, email, social or direct audience-building.

None of those practices is spectacular by itself. The value is in doing them consistently, learning from the site and resisting the temptation to rebuild everything every time one metric moves.

The first adaptation: answer the question faster

One of the earliest changes was structural. Long articles that made readers work too hard to find the answer were a problem for humans and for AI systems trying to summarize the page.

I began pushing toward a simpler pattern:

The second adaptation: stop treating every page as an island

I had a lot of content. That was not the same thing as having a clear system.

The more I reviewed the site, the more I found pages that were useful individually but weak as part of a journey. So I began rebuilding around hubs, series navigation, reverse links and clearer next-step routes.

That work matters for SEO, but it also matters for AI systems. If several pages all discuss similar things without making their relationship explicit, the machine has to infer the hierarchy. I would rather tell it.

The third adaptation: product names had to make sense outside my own head

Creator businesses often use internal names that loyal followers understand but first-time visitors do not.

That becomes more dangerous when an AI system is trying to decide whether a product fits a user’s request.

I started tightening product titles, summaries and access explanations so a person or agent could answer basic questions without knowing the Jack Righteous backstory first:

• What is this?
• Who is it for?
• What does it include?
• What problem does it solve?
• What happens after purchase?
• How is it different from the next product?

The fourth adaptation: I stopped treating AI SEO as a separate trick

AI SEO is useful language because it forces us to notice that ChatGPT, Gemini, Perplexity, Claude and other systems are becoming discovery layers.

But I no longer think the goal is to produce “AI-optimized copy.”

The better goal is to make the business legible.

If the content is accurate, the product records are explicit, the internal relationships are clear, and the evidence is easy to trace, the same work helps people, search engines and AI systems.

That is much more durable than chasing one markup trick or one prompt-monitoring score.

Kedra helped expose the gap between the business I saw and the business AI saw

When I started testing AI visibility more deliberately, one of the most useful lessons was not a ranking. It was the gap between the areas I believed Jack Righteous was known for and the areas AI systems appeared to recognize most clearly.

That is actionable intelligence.

If an AI system consistently understands the business as “creator resources” but fails to connect it strongly enough to “AI music training,” I can examine whether the problem is missing content, weak internal linking, vague product language, insufficient evidence or simply a category where other sites currently have stronger authority.

The point is not to argue with the machine. The point is to understand the signal and decide whether the site needs to change.

Then the agents started contacting me

This is where the experiment moved beyond AI search.

In my public AI Is Making Contact — Log 01 and Log 02, I document AI agents that approached me as creators, submitted work and asked for the same kind of judgment and direction human creators ask me for.

That changed the question again.

Shopify is making the same question operational

Shopify’s agentic-commerce work now connects discovery, product data, carts and supported checkout experiences. That means the journey is beginning to change from:

AI recommends → human clicks → website does everything

toward:

AI discovers → AI compares → human approves → agent or embedded commerce flow moves toward checkout

That is why I published the public analysis Shopify Opens Checkout to AI Agents. Amazon Draws the Line.

The public article explains the platform fight. This member article explains what I am changing inside my own business because of it.

The model I am working toward: three customer interfaces

What I am changing now

• Clearer titles and summaries: less clever ambiguity, more direct description of the problem solved.
• Stronger hubs: major topics need an obvious source-of-truth page and a visible relationship to supporting articles.
• Reverse linking: new reporting should update older evergreen guides, not only link one way.
• Product differentiation: free, one-time library, consultation and deeper-access offers need to be distinguishable without insider knowledge.
• Explicit support boundaries: an agent should be able to understand when it can proceed and when its human needs to step in.
• Better evidence loops: I am prioritizing source, landing page, next click, signup, purchase and support behavior over vanity traffic.
• Agent-facing thinking: I am beginning to ask whether the same information a human needs is structured clearly enough that an authorized agent could use it safely.

What I am not doing

• I am not abandoning Google SEO.
• I am not assuming every AI referral is valuable.
• I am not redesigning the site around hypothetical autonomous buyers.
• I am not treating an AI agent as automatically authorized because it claims to represent someone.
• I am not giving sensitive customer information to a bot merely because it asks.
• I am not assuming Shopify, ChatGPT, Gemini, Muse or any other platform will guarantee visibility or sales.

My next experiment

The next stage is not simply measuring whether an AI mentions Jack Righteous.

I want to follow the chain:

At the same time, I am continuing the creator-side experiment documented in AI Is Making Contact: can AI creators use the same development system I use with humans, and does the advice change what they actually do next?

Those two experiments are beginning to converge.

Related JR learning path

This case study is part of a larger set of free and member resources. If one part of the process is new to you, use the relevant guide rather than trying to learn everything at once.

Member next steps

If you are building a creator business, choose one route rather than trying to copy everything I changed at once.

Shopify for Creators Agentic Commerce Guide AI Is Making Contact — Log 02

Access paths

This case study is part of the ASK JACK Member Resource Library. Current access routes include:

• JR Creator Library — One-Time
• Creator Pro
• Complete Access

Use the level that matches how much support you actually need. The point is not to buy the largest package. The point is to have the right resource when the next decision becomes real.

Search visibility, referral attribution and AI-shopping features change over time. Treat this as a working method rather than a fixed formula: make useful changes, measure over meaningful periods, keep what helps and continue adjusting.