Metrisque launches AI visibility measurement tool

Metrisque has launched a measurement instrument that claims to show where AI models place a brand, product or service and to return the same answer each time it is queried. The company says its pre-registered study found 92% of AI recommendations landed where the instrument predicted before the models were asked.

Why it matters: - Metrisque is targeting a basic problem in AI marketing measurement: large language models can change answers from one run to the next, making visibility hard to prove. - The company says its instrument gives companies a consistent reading on whether their messaging matches what buyers ask AI assistants. - Metrisque is positioning AI answers as a channel that now needs a measurement layer, similar to search and social.

What happened: - Metrisque launched a tool that measures where AI models place a brand, product or service. - The launch was announced Sept. 6, 2026, in San Francisco. - The company says the instrument returns the same answer every time it is queried in an LLM. - Metrisque also published a pre-registered study with a permanent citation, DOI 10.5281/zenodo.21417361. - The study tested roughly 1,100 real recommendations from two leading AI models against predictions made before the models were asked anything. - The company says 92% of the recommendations landed where the prediction said they would. - Co-founder Lian Pham said: “We also published the prediction we got wrong. A company that only shows you its wins hasn't shown you anything.”

The details: - Brand Recall measures what AI models already believe about a company before a user asks a question. - Category Fit measures how AI models classify a company and its individual products. - Buyer Match measures whether a company’s wording matches the questions buyers ask AI assistants. - AI Recommendations shows which brands get named when a buyer asks and where a company ranks against them. - Competitor's Citations shows which websites each AI model reads when answering a buyer’s question and which of those sites already name the company. - Question Finder is designed to identify which buyer question is worth pursuing and whether that question is open, contested or already owned. - Metrisque says its research found that three leading AI models often read almost entirely different websites for the same buyer question. - On one question, the company says 46 sources were cited, only one was cited by all three models, and 39 were cited by a single model. - Metrisque says the same model asked the same question days apart changed roughly 60% of the websites it read.

Between the lines: - The product is built around the idea that AI visibility should be measured by match quality and model behavior, not just by counting brand mentions. - The study’s pre-registered design and public citation are meant to support credibility in a market where model answers can be volatile. - Pham framed the need as a budget and accountability issue, saying every major channel got a measurement layer before it got funding. - A pilot customer, Rosmon Sidhik, said two of three AIs did not know the company and the one that did miscategorized it, and that changing one page moved the measurement.

What's next: - Metrisque is available at the company’s website. - Companies can calibrate their visibility posture for brands, products and services. - The company also pointed readers to its LinkedIn page at Metrisque on LinkedIn.

The bottom line: - Metrisque is betting that AI answer quality can be measured, tracked and improved the same way marketers track search and social performance.