SEO, GEO and AEO consultants: we're looking for partners. See the program →
Metrisque
AI visibility, measured before the answer

Get an AI visibility diagnosis and correction plan

See how AI understands your brand. Investigate the gaps and get specific changes to make.

Start free diagnostic See a sample diagnosis
ASSISTANTS COVERED BY THE SELECTED PLAN
ChatGPTGeminiClaudePerplexity (coming soon)Grok (coming soon)Muse Spark (coming soon)

Free plans use Gemini. Category Fit uses ChatGPT and Gemini. Grey logos are not measured yet.

READS THE SAME CATALOG
Meta MuseGoogle AI ModeMicrosoft CopilotPerplexity

Agents that discover products through Shopify Catalog and the Universal Commerce Protocol. The same product data we measure.

I run one business
Start with your own website.
Run a free diagnostic and review the findings before choosing a paid plan.
I'm an agency or partner
Add diagnosis to your agency workflow.
Use Metrisque alongside your tracker to investigate client gaps and plan corrections.
What you get

Your report connects each finding to an action.

Buyer question → finding → evidence → suggested correction → re-test.

Each finding shows its evidence, so you can decide what to change.

From our pre-registered study

What our study measured, and its limits.

92%
of recommendations in memory-only tests came from inside the territory we measured
1,916
products measured
104,198
published texts
3
assistants: ChatGPT, Gemini, Claude

The study did not measure traffic, buyer behavior, or revenue. See the study, its scope and its limits →

The Metrisque Rings™ · patent pending

The Gartner Magic Quadrant for AI answers.

The Metrisque Rings place a page around the buyer's question by how closely its meaning matches.

  • Ring 1 is a strong measured match.
  • Ring 2 is a partial measured match.
  • Outside is a weak measured match.

A strong match does not guarantee a future recommendation.

OUTSIDERING 2RING 1You
What we measure

Choose the measure for your next task.

1
Brand Recall
Check what AI says about your brand and which facts it gets wrong.
2
Category Fit
Check how AI classifies your product pages.
3
Question Finder
Compare answers before choosing buyer questions to work on.
4
Buyer Match
Check how directly a page answers a buyer question.
5
Sources
See the pages cited in answers and the brands they mention.
6
AI Recommendations
Record the brands recommended in the tested answers.
7
Content
Review specific page edits and the facts they need.
See it on one brand · illustrative example

Here's Marlowe Vance.

24 of 180 products are classified outside their category.
A setting spray is classified as workwear.
Category Fit · Marlowe Vance · illustrative
118 in their category
38 close
24 outside their category
Night Shift Setting Spray
Makeup › Setting Spray
AI classifies it as Workwear
On “best long-wear lipstick”, Marlowe Vance is not recommended.
Its page is a partial match, in Ring 2. The recommended brands are strong matches.
The Metrisque Rings™ · illustrative
“best long-wear lipstick”
OUTSIDERING 2RING 1Halden BeautyKestrel CosmeticsMarlowe Vance
You
Recommended by AI
Considered
Content marks the lines that weaken the match, and keeps the useful facts.
Content · Marlowe Vance · illustrative
Underlined helps
Shaded weakens the match

Born after midnight, made for the after-party. A matte lipstick that stays on for 12 hours without drying your lips. Your mood, in a tube.

Correction
Born after midnight, made for the after-party.
A long-wear matte lipstick that lasts 12 hours without drying your lips.
How it works

Diagnose, cause, correction.

1
Diagnose
Check your brand facts, product categories, page relevance, citations, and recommendations.
2
Cause
Review the evidence behind incorrect facts, mismatched categories, and missing answers.
3
Correction
Get proposed changes for each finding. Publish the edits you approve, then re-test against your baseline.
Your next step

Start with your own website.

Run a free diagnostic and review the findings before choosing a paid plan.

Start free diagnostic See a sample diagnosis
Measured, not guessed.