Signal Harbor Weekly: June 11, 2026

The Invisible Rankings

Hey,

Welcome to the first issue of Signal Harbor Weekly.

Every week, I'll send one short email about what's actually happening in AI search and what it could mean for your business.

THE TEST

Most businesses know where they rank on Google.

They know how much traffic their website receives.

They know which advertisements generate clicks and which pages produce leads.

But almost none of them know how they appear inside ChatGPT, Claude, Gemini, or Perplexity.

So I ran a simple test.

I asked two AI platforms the exact same question:

"Who are the best roofing companies in Chicago?"

They produced two different lists.

Different companies.

Different sources.

Different explanations.

Some businesses appeared in one answer and disappeared from the other.

Some were supported by strong third-party sources.

Others were mentioned briefly or not at all.

And the companies being evaluated would have almost no visibility into why they were selected, how they were described, or why a competitor appeared instead.

THESE ARE THE INVISIBLE RANKINGS

The rankings that increasingly shape who gets discovered, trusted, and contacted before a buyer ever visits a website.

AI search is not one universal ranking system.

ChatGPT may recommend one group of companies.

Claude may recommend another.

Gemini and Perplexity may rely on different sources, emphasize different factors, and construct completely different answers.

A business can therefore appear highly visible on one platform and nearly invisible on another.

It can also be:

• Placed in the wrong category

• Described using outdated information

• Reduced to a generic list item

• Misrepresented on an important feature

• Supported by weak or irrelevant sources

• Left out in favor of a less qualified competitor

This is not simply a rankings problem.

It is a measurement problem.

Businesses can answer questions such as:

• Where do we rank on Google?

• How many people visited our website?

• Which campaign generated a lead?

• Which landing pages converted?

But most cannot answer:

• How often does AI recommend us?

• Which competitors appear instead?

• Which sources are shaping the answers?

• Is the information being presented about us accurate?

• Are our strongest advantages actually influencing the response?

• How much does performance change across platforms?

Most companies are entering the AI-discovery era without a clear way to measure any of this.

They are operating blind.

THE REAL CHALLENGE

A single ChatGPT screenshot is not an AI visibility strategy.

AI answers naturally change from one search to the next.

A company might appear first in one response, disappear in another, and be replaced by a competitor later that day.

That means a one-time audit can create false confidence.

If an agency runs one prompt and presents the result as your definitive AI ranking, it is not measuring your true market position.

It is showing you one moment inside a constantly changing system.

One screenshot is an anecdote.

Measurement requires repetition.

But visibility alone isn't enough.

There is a major difference between being included in an answer and actually shaping it.

A company might be cited at the bottom of a response while the AI uses competitor information to explain the category, define the buyer's needs, and make the recommendation.

Being present is not the same as being influential.

The goal is not simply to get cited.

The goal is to shape the answer.

THE BUSINESS RISK

Imagine a potential customer asks:

"What is the best construction-management software for a residential remodeling company?"

One company may have the strongest product for that customer.

But the AI system could still:

• Recommend a competitor first

• Misunderstand its ideal customer

• Rely on an outdated comparison article

• Omit an important capability

• Cite the company without using its strongest evidence

• Position the product for the wrong type of buyer

• Leave the company out entirely

The buyer may never open Google.

They may never review ten search results.

They may accept the AI-generated shortlist and move directly into product comparisons or demo requests.

The commercial decision can begin taking shape before the buyer ever reaches the company's website.

Traditional analytics measure what happens after the visit.

AI intelligence must measure what influenced the buyer before it.

WHAT SIGNAL HARBOR DOES

Signal Harbor provides an AI intelligence and diagnostic layer for the AI discovery era.

We measure how companies are represented, recommended, and understood across ChatGPT, Claude, Gemini, Perplexity, and other generative-search platforms.

Our system evaluates:

• Recommendation frequency

• Competitive share of recommendation

• Citation influence

• Source patterns

• Cross-platform differences

• Category-level performance

• Inaccurate and misleading claims

• Visibility changes over time

We run the repeated testing and structured analysis required to separate real competitive patterns from random output changes.

Then we diagnose why those patterns exist.

The objective is not to create another marketing dashboard.

It is to give leadership, product, content, and growth teams a clear view of how AI systems interpret their company and a blueprint for improving that interpretation.

Businesses spent decades building infrastructure to measure Google, advertising, and website behavior.

Now they need infrastructure to measure AI.

And most companies have not started.

WANT TO SEE YOUR INVISIBLE RANKINGS?

Reply with:

• Your company name

• Your website

• The most important question your buyers ask

I'll select a small number of companies to participate in a future AI visibility benchmark and share how the major AI platforms currently represent their category.

Sebastian Miller
Founder, Signal Harbor
signalharborconsulting.com
sebastian.miller@signalharborconsulting.com
402-306-2213

Research Note

The ideas in this issue are informed by emerging research into generative-search variability, citation behavior, and repeated AI measurement, along with Signal Harbor's own cross-platform testing. As this field develops, Signal Harbor will continue evaluating new research and refining how AI recommendation performance should be measured.

Signal Harbor Weekly helps business leaders understand how AI systems represent, recommend, and influence buying decisions.