R-U-Known Methodology

How R-U-Known Measures AI Visibility: From Benchmark Evidence to Action

AI visibility is more complicated than asking an AI system one question and checking whether your company appears in the answer.

In short: R-U-Known is an AI visibility tracker for professionals, busy executives who prefer a specialist, and agencies tracking visibility for clients. It measures how AI finds, mentions and recommends brands, then turns the evidence into practical actions to improve visibility.

Last updated: October 2, 2026 • Author: R-U-Known Team • Category: Methodology

People ask different questions, use different language and look for different capabilities. AI systems may know some things about a business very well while failing to connect that same business with other services, locations, features or buying situations.

R-U-Known was designed to measure those differences systematically.

Its methodology follows a repeatable process:

Questions → AI Responses → Sources → Evidence → Analysis → Frozen Benchmark → Action Plan → Changes → New Benchmark

The purpose is not simply to produce an AI visibility score. It is to preserve enough evidence to understand what happened, investigate why it happened and identify what a business can do next.

What Does AI Visibility Actually Mean?

When someone searches using a traditional search engine, visibility can often be examined through rankings, impressions, clicks and search positions.

AI search behaves differently.

A user may ask:

Which companies provide a particular service?
Who would you recommend for a specific requirement?
Which businesses operate in a certain area?
Does a particular company support a specific capability?
What are the alternatives to a particular product?
Which provider is suitable for a particular type of customer?
How do several companies compare?

A company may appear in one answer and be completely absent from another, even when both questions are relevant to what that company does.

That is why R-U-Known does not treat AI visibility as a single search result.

Step 1: Building a Structured AI Visibility Benchmark

R-U-Known measures a company across a structured set of questions.

These questions are intended to examine the business from multiple discovery angles rather than repeatedly asking variations of the same question.

Depending on the company and market, questions can explore areas such as:

Brand discovery
Products and services
Capabilities
Geographic relevance
Customer requirements
Category discovery
Competitive discovery
Recommendation scenarios
Specific buying or operational requirements

Together, these questions form a benchmark.

A benchmark is a point-in-time measurement of how the selected AI search environment responds to those questions at the time the benchmark is run.

This is important because AI visibility is not static. Sources change, websites change, AI systems change and the information available to those systems changes over time.

Step 2: Capturing the Actual AI Responses

For each benchmark question, R-U-Known records the response returned by the AI search process.

The first obvious question is:

Was the target company discovered?

But R-U-Known does not stop there.

A brand mention alone cannot explain the complete result.

The surrounding answer can reveal how the business is understood, which capabilities are associated with it, which competitors appear alongside it and which relevant concepts the AI system does not appear to connect with the company.

This gives R-U-Known something more useful than a simple mention counter: context.

Step 3: Preserving the Sources Behind the Answer

Where source or grounding information is available from the AI search process, R-U-Known preserves that information alongside the response.

This allows the analysis to investigate not only what the AI said, but also which pages formed part of the available evidence surrounding that answer.

This distinction matters.

A visibility score tells you what happened. Evidence helps explain why.

For example, an AI response might mention several businesses while relying on category articles, directories, official company pages, discussion pages or other web sources.

Understanding those sources provides valuable context about the information environment in which the answer was produced.

Step 4: Not Every Source Means the Same Thing

One of the important principles in R-U-Known's methodology is that the presence of a URL in the source evidence does not automatically mean that the page is evidence about the company being measured.

A page may instead provide general information about the category, describe a competitor or simply help establish context for the AI response.

R-U-Known therefore analyses the role that sources play in the evidence.

Target-brand evidence

Evidence that directly relates to the company, brand or product being measured.

Competitor evidence

Evidence that relates primarily to another company or product appearing in the same discovery environment.

Category or contextual evidence

Evidence that helps establish the wider topic, category or question context without necessarily providing direct evidence about the target company.

This distinction prevents a source from being treated as target evidence merely because it happened to appear among the pages associated with an answer.

Step 5: Analysing Visibility in Context

Once the responses and evidence have been collected, R-U-Known analyses the benchmark to identify patterns.

The purpose is not merely to answer:

“How many times was the company mentioned?”

It is also to investigate questions such as:

Where is the company consistently discovered?
Where is it absent?
Which capabilities are clearly associated with it?
Which capabilities appear weakly understood?
Which competitors appear in relevant answers?
What sources repeatedly influence the discovery environment?
Where does the available evidence support the target company?
Where is the evidence primarily about competitors or the wider category?
Where does the AI response appear uncertain or incomplete?

These patterns provide the intelligence layer between raw AI responses and an actionable improvement plan.

Step 6: Freezing the Benchmark

When a benchmark is completed, it represents what was observed at that particular point in time.

R-U-Known treats that benchmark as immutable.

R-U-Known does not rewrite history.

Later analysis may help users understand the evidence more deeply, but the original benchmark remains the reference point against which later measurements can be compared.

This separation between measurement and interpretation is important.

If historical benchmarks were silently changed whenever the analysis evolved, it would become difficult to determine whether AI visibility had genuinely changed or whether the underlying measurement had simply been rewritten.

The evidence stays fixed. The understanding of that evidence can become deeper.

Step 7: Turning Evidence Into an Action Plan

Measurement is useful, but eventually every business reaches the same question:

“What do we do about it?”

This is where R-U-Known's Action methodology begins.

A user selects a completed frozen benchmark and R-U-Known uses the evidence and analysis from that benchmark to identify areas that may deserve attention.

These can include patterns such as:

Important capabilities that AI does not appear to associate clearly with the company
Relevant questions where competitors appear but the target does not
Missing or weak supporting information
Ambiguous product or service messaging
Important business information that AI cannot establish confidently
Content gaps around relevant customer questions
Weak connections between existing capabilities
Source patterns that reveal where the discovery environment is being shaped

R-U-Known then turns those findings into an improvement plan.

Recommendations Are Connected Back to Evidence

An Action plan should not simply produce generic advice such as “write more content” or “improve your SEO.”

R-U-Known is designed to connect recommendations back to the benchmark evidence that caused the issue to be identified.

That allows the user to investigate:

Which benchmark questions exposed the issue
What the AI response actually said
Whether the target company appeared
Which competitors were present
Which sources were associated with the answer
Why the recommendation was generated

This creates a much more useful relationship between measurement and action:

Finding → Evidence → Explanation → Recommended Action

R-U-Known Does Not Claim to Control AI Answers

No website owner can directly control what an independent AI system will say or guarantee that a particular change will cause a brand to be recommended.

AI responses can be influenced by many factors, including available web information, source selection, model behaviour, question wording and changes to the AI system itself.

R-U-Known therefore does not treat an Action recommendation as a guaranteed ranking formula.

Instead, the purpose of the Action plan is to identify evidence-backed areas that a business can improve, clarify or strengthen.

The effect of those changes can then be measured in a later benchmark.

Step 8: Make the Changes and Measure Again

This is where the methodology becomes a cycle rather than a once-off report.

After reviewing the Action plan, a business may decide to improve website content, clarify product positioning, explain capabilities more explicitly, strengthen important pages or address other evidence-backed gaps.

The next benchmark provides a new point-in-time measurement.

The process becomes:

Benchmark 1 → Evidence → Analysis → Action → Changes → Benchmark 2 → Compare

Instead of assuming that the changes worked, the business can observe what happened in the next measurement.

Why Historical Benchmarks Matter

A single benchmark is a snapshot.

Multiple benchmarks create a history.

Once measurements are preserved over time, businesses can begin asking more useful questions:

Is our AI visibility improving?
Are more relevant questions discovering us?
Are previously misunderstood capabilities becoming clearer?
Are different sources beginning to support our visibility?
Has our competitive position within AI answers changed?
Did the areas we worked on show movement in later benchmarks?

Historical comparison turns AI visibility from a momentary score into something that can be observed over time.

Why R-U-Known Separates Measurement, Evidence and Action

These three concepts serve different purposes.

Measurement answers:

What happened?

Was the company discovered? How frequently did it appear? Where was it absent?

Evidence answers:

What information surrounded that result?

What did the AI say? Which pages appeared in the source environment? Which sources relate to the target, competitors or the wider category?

Action answers:

What can we do next?

Which evidence-backed gaps can reasonably be addressed through clearer information, stronger content, better positioning or other improvements?

Keeping these layers connected but distinct helps prevent interpretation from being mistaken for raw measurement.

From AI Visibility Scores to AI Visibility Intelligence

A percentage can be useful.

But a percentage on its own cannot tell a business why it is visible, why it is missing, what information AI is relying on or what should be investigated next.

R-U-Known therefore treats the score as the beginning of the investigation, not the end.

The deeper value comes from connecting:

Visibility + Context + Competitors + Sources + Evidence + History + Action

That combination turns a measurement into intelligence a business can use.

A Repeatable Process for Improving AI Visibility

AI discovery is still evolving rapidly.

R-U-Known does not attempt to reduce that complexity to a magic formula. Instead, it provides a repeatable methodology for observing it.

Measure how AI currently discovers the business.
Preserve the responses and available evidence.
Analyse visibility, context, competitors and sources.
Freeze the benchmark as a historical point-in-time record.
Identify evidence-backed gaps and opportunities.
Act on the areas the business decides are important.
Measure again with a new benchmark.
Compare what changed.

This turns AI visibility from something businesses can only guess about into something they can observe, investigate and systematically work to improve.

Frequently Asked Questions

What does R-U-Known measure?

R-U-Known measures how a company, brand or product appears across a structured benchmark of relevant AI search questions. It examines visibility in context rather than relying on a single prompt or isolated brand mention.

Does R-U-Known only count brand mentions?

No. Brand discovery is an important measurement, but R-U-Known also preserves AI responses and available source evidence so users can investigate the context surrounding visibility and non-visibility.

Why does R-U-Known use multiple questions?

AI systems may understand a company differently depending on how a question is asked. Testing multiple relevant discovery scenarios provides a broader picture than relying on a single prompt.

What is a benchmark in R-U-Known?

A benchmark is a structured point-in-time measurement consisting of AI search questions, responses and the associated evidence captured during that run.

Why are completed benchmarks frozen?

A completed benchmark represents what was observed at a specific point in time. Keeping it immutable preserves historical integrity and allows future benchmarks to be compared against the original measurement.

What is evidence in R-U-Known?

Evidence includes the information captured around an AI response, including available source or grounding pages and the context required to investigate how the answer was formed.

Does every source count as evidence about my company?

No. A source may relate directly to the target company, to a competitor or simply to the broader category or topic. R-U-Known distinguishes these roles rather than assuming every source is target-brand evidence.

What is the R-U-Known Action plan?

Action uses a selected completed benchmark to identify evidence-backed gaps, unclear messaging, missing information and other areas that may be limiting AI visibility. It then turns those findings into recommendations the user can investigate and act on.

Are Action recommendations generic AI advice?

The purpose of Action is to derive recommendations from the selected benchmark and connect them back to the evidence that produced the finding, rather than generating an unrelated generic checklist.

Does following an Action recommendation guarantee better AI visibility?

No. Independent AI systems determine their own responses, and many factors can influence those answers. R-U-Known identifies evidence-backed areas that may be improved and allows the effect of changes to be observed through later benchmarks.

How do I know whether the changes worked?

After changes have been made, a new benchmark provides another point-in-time measurement. Comparing benchmarks allows users to investigate whether visibility, associations, sources and other measured signals changed over time.