How tucii works
tucii checks whether AI answer engines discover, recognize and consider a public company, brand, product, organization or professional. You get a point-in-time score, with the questions, answers and calculation behind it.
From a name to the evidence
Find the right match
Enter a public name or website. tucii uses public information to identify the target and what it offers. Review the match and correct it if needed.
Approve the scan focus
Check the needs and discovery themes before starting measurement. The discovery questions describe relevant problems and situations without naming the target. Separate named questions check recognition and consideration.
Read what came back
The report shows the score, engine coverage, exact stored questions, answers and returned citations. Open the evidence to understand a result. Missing or unresolved checks remain visible.
What does the score measure?
The current AI Discovery Score combines three components on a 0–100 scale. Discovery is unprompted by the target’s name; recognition and consideration explicitly name it.
- 50%Discovery
- Does it appear in answers to questions that do not name it?
- 20%Recognition
- When named, does the answer identify what it is and what it does?
- 30%Consideration
- When named, does the answer include it on a shortlist for a relevant need?
For each qualifying engine, tucii adds the three weighted component points. It then averages those engine scores with equal weight and rounds once. Consideration weights the two tested needs equally. Missing checks are excluded; their weight is not moved to another component.
Discovery Rate is the simpler measure: the percentage of eligible unbranded answers that substantively mention the target. It is separate from the combined score. Appearing only in a citation does not count as discovery.
The score is a provisional index. It is not an SEO score, a search ranking, a business-quality rating or a market percentile.
Which AI engines are checked?
- ChatGPT
- Google AI Mode
- Microsoft Copilot
- Gemini
The current plan sends 21 English questions to each engine: 15 discovery, 2 recognition and 4 consideration questions. That is 84 planned answers, with one observation per question and engine. The report identifies which engines and checks actually contributed.
Answers are observed through a managed collector on the named consumer AI services. They can differ from your own session. tucii does not substitute an official model API answer when a consumer observation fails.
A missing answer is not a miss
An engine qualifies only with at least 12 of 15 resolved discovery checks, both recognition checks, and at least 3 of 4 consideration checks covering both needs. At least two engines must qualify for an overall score.
A partial report can show a score when enough evidence qualifies, while identifying missing checks and excluded engines. If fewer than two engines qualify, the report shows insufficient data instead of inventing a zero.
A low score is different: completed, usable answers may simply not mention or shortlist the target. The report separates this observed result from unavailable evidence.
What a snapshot can tell you
It shows what happened for this set of questions at the recorded time. AI answers vary with time, market, session and product behavior. A single scan does not establish a trend or explain what caused the result.
The score describes observed answer behavior. Citations and accuracy warnings help you inspect claims; they do not prove that every answer is correct. Historical reports keep the methodology used when they were created.
Scanning does not change a website, train an AI model or guarantee better visibility. Read the FAQ for access, timing and sharing details.