Find the meaning.
Follow the signal.
The terms behind your AI visibility dashboard, explained in plain language.
18 terms to help you read the evidence
Read the measurements
Understand the numbers before drawing a conclusion.
- AI visibility
How a brand appears in the AI answers you observe. In Echuno, visibility is interpreted through saved answers, brand mentions, recommendation positions, and sources within a selected scope. It is not a count of every conversation happening on a provider.
Read the practical guide for AI visibility- Brand mention rate
Answers mentioning a brand divided by answers with a recorded assessment of that brand, expressed as a percentage. Answers without a brand assessment are excluded from this calculation.
Example8 mentions in 16 assessed answers gives a 50% mention rate, even if another 4 saved answers are still unassessed.
Read the practical guide for Brand mention rate- Recommendation rank
A recorded position in an explicitly ordered set of product recommendations. A mention can have no explicit rank. Average rank uses only recorded positions; lower numbers describe earlier recommendations in those answers.
ExampleRecorded ranks of 1, 2, and 3 average to 2.0. An additional unranked mention does not enter that average.
Read the practical guide for Recommendation rank- Collection coverage
Which planned questions and providers have saved answers in the reporting period. A question with one saved ChatGPT answer may be sampled while still missing Claude and Gemini answers. Coverage and brand mention rate describe different things.
Read the practical guide for Collection coverage- Percentage point
The arithmetic difference between two percentages, often abbreviated pp. It differs from a relative percentage change.
ExampleA rate moving from 40% to 50% increased by 10 percentage points. Relative to 40%, that is a 25% increase.
Read the practical guide for Percentage point
Understand collection
Know what was asked, saved, and still waiting.
- Buyer question
A prompt that reflects something a potential buyer might ask AI while exploring a problem, comparing options, or choosing a product. Keep core wording stable when comparing saved answers over time.
Example“Which tools help a small marketing team track brand mentions in AI answers?”
Read the practical guide for Buyer question- Enabled question
An unarchived question eligible for configured automatic collection. Enabling a question does not turn on a daily schedule or immediately create a provider request. The schedule, selected providers, allowances, and existing attempts still affect collection.
Read the practical guide for Enabled question- Paused question
An unarchived question excluded from future automatic collection. Its saved answers remain available, and manual collection is still possible. Pausing a question does not cancel browser requests that are already queued or running. Enabled and paused questions both count toward the project question limit.
Read the practical guide for Paused question- Archived question
A question removed from the current question library and excluded from new collection jobs. Its saved answers remain in the answer library. Archived questions can be restored as paused questions when there is room within the project limit. Existing browser requests are managed separately.
Read the practical guide for Archived question- Saved answer
A captured response with its question, provider, collection time, and available context. Saving an answer and assessing its brand mentions are separate stages. A saved response can still be awaiting analysis.
Read the practical guide for Saved answer- Temporary chat
A provider’s temporary or incognito conversation mode. It helps reduce the influence of earlier conversation history. It does not prove that account settings, location, model version, or other conditions were identical across captures. Echuno labels this mode only when it was recorded.
Read the practical guide for Temporary chat- Queue wait
Time between queuing a provider request and starting collection. It can include waiting for the browser runner, earlier requests, an available provider session, or a maintenance pause. It is separate from answer-generation or collection time.
Read the practical guide for Queue wait- Collection duration
Elapsed time from starting a provider request to saving its answer. It includes the collection workflow and should not be described as pure model-generation latency. Queue-to-save duration also includes queue wait.
ExampleA 30-second collection after 4 minutes in the queue takes 4 minutes 30 seconds from queueing to saving.
Read the practical guide for Collection duration
Follow the evidence
Connect a result to the words and sources behind it.
- Brand mention
An answer identifying a tracked brand through its name, recorded alias, or domain. A mention does not automatically mean the answer recommends the brand, links to its website, or gives it a rank.
Read the practical guide for Brand mention- Brand assessment
A recorded analysis of one brand in a saved answer, including whether it was mentioned and any supported position or evidence. An absent assessment is unknown; it is not a confirmed absence of the brand.
Read the practical guide for Brand assessment- Evidence quote
An exact excerpt supporting a recorded brand assessment. Read the surrounding saved answer to understand the claim, especially when a product is named incidentally or compared with alternatives.
Read the practical guide for Evidence quote- Citation
A source reference recorded with an answer. A citation can support a claim without recommending a product. A brand can also be mentioned without a citation to its own website.
Read the practical guide for Citation- Source domain
The website hosting a recorded source, such as example.com. Grouping source links by domain helps identify recurring websites in your saved answers. Frequency in your sample does not establish why a provider recommended a brand.
Read the practical guide for Source domain
Your next insight starts
with a better question.
Add your brand. Choose your questions. See what AI says.