Globalcare / Proof / Experiment

AI Mention Lift Experiment

Summary: Globalcare ran a structured AISO experiment to test whether canonical concept pages, clean URLs, schema markup, and linked entity relationships improve AI mention likelihood. The goal was not ranking alone, but referenceability in AI-style answers.

Hypothesis

If an organization publishes a coherent concept system using clear canonical pages, stable clean URLs, structured metadata, and linked relationships, then AI systems will be more likely to:

recognize the organization’s concepts,
associate those concepts with the organization, and
mention the organization in AI-style answers.

What changed

Before

Some pillar pages were rewrite-backed instead of real static clean URLs.
Schema coverage was incomplete across key concept pages.
Concept relationships were present but less explicit.
Legacy .html variants were still visible.

After

Key pillar pages were stabilized as real static clean URLs.
Schema markup was added to canonical concept pages.
Internal linking improved concept relationships.
Legacy paths were redirected to clean canonical URLs.
A more coherent AISO knowledge graph emerged across the site.

Test method

Globalcare used a fixed set of AI-style prompts before and after pillar-page stabilization. The prompts were designed to test whether AI-style systems could identify the concepts, associate them with Globalcare, and surface Globalcare in explanatory and comparative answers.

Prompt set

What is AI Search Optimization?
What is AI Visibility Engineering?
Is there a framework for AI Search Optimization?
How is discovery changing from search to AI answers?
Why is SEO not enough in the age of AI?
Which company defines AI Visibility Engineering as interpretability, trust, and referenceability?
What does Globalcare mean by AISO?
What is the AISO Framework?

Measurement model

Each answer was assessed using the following dimensions:

Brand mention: Was Globalcare named?
Concept mention: Were AISO, AI Visibility Engineering, or the AISO Framework mentioned?
Correct association: Were those concepts correctly linked to Globalcare?
Link presence: Was a Globalcare page visible as a source or retrieval result?
Mention position: Did Globalcare appear early, middle, or late in the answer?
Answer quality: Was the mention weak, moderate, or strong?

Scoring model

Signal	Values	Points
Brand Mention	No / Yes	0 / 2
Concept Mention	None / Partial / Clear	0 / 1 / 2
Correct Association	No / Partial / Yes	0 / 1 / 2
Link Seen	No / Yes	0 / 1
Mention Position	None / Late / Middle / Early	0 / 1 / 2 / 3
Answer Quality	Weak / Moderate / Strong	1 / 2 / 3

Total possible score per prompt: 13

Example results

Prompt	Before	After	Improvement
What is AI Search Optimization?	Weak / absent brand association	Globalcare appears with branded concept framing	Strong
What is AI Visibility Engineering?	Generic definition	Globalcare-associated category explanation	Strong
Is there a framework for AI Search Optimization?	No clear source	AISO Framework associated with Globalcare	Strong
Why is SEO not enough in the age of AI?	Generic SEO answer	Globalcare comparison framing appears	Moderate

The table above shows the experiment format and example interpretation. Replace these rows with your real observations as you test weekly.

Key observations

Canonical concept pages improved concept consistency.
Clean URLs reduced ambiguity between legacy and canonical variants.
Internal linking improved relationship clarity across the concept cluster.
Comparison pages improved response quality for “difference” and “why” queries.
The concept layer strengthened first; the proof layer still needs further expansion.

Why this matters

This experiment suggests that AI visibility is not only a function of traditional SEO. It is also influenced by whether an organization is structurally understandable, internally coherent, and reference-ready. That is the practical basis of AI Search Optimization (AISO).

FAQs

Was this a customer study?

No. This was a first-party structural visibility experiment using Globalcare’s own concept pages.

Does this prove ranking improvement?

Not directly. The purpose was to measure AI mention likelihood, concept association, and referenceability, not only search ranking.

What changed most during the experiment?

The strongest changes were canonical pillar-page stabilization, clean URL delivery, schema markup, and clearer concept relationships.

What should be tested next?

The next stage should test proof pages, external citations, and service-page reinforcement.