What is MCAC?
MCAC is LS-Labs' operational diagnostic framework for examining brand Mention, source Citation, factual Accuracy and Conversion-readiness in AI answers.
These are LS-Labs working definitions. They are not a search platform's published ranking standard, and the four dimensions should not be collapsed into a universal brand score.
The four checks
| Dimension | What is checked | Important boundary |
|---|---|---|
| Mention | A brand name or clearly identifiable reference in the answer's body | A mention can recommend, compare or criticize. Brand-named questions must be identified separately. |
| Citation | Visible links, footnotes or source lists and whether they support the associated claim | Visible citations do not reveal every source or influence behind a model's output. |
| Accuracy | Whether facts such as address, phone number and price agree with a valid reference | Correct, incorrect, not provided and unable to verify are different states. |
| Conversion-readiness | Next-step information such as a phone number, website or address | An available action route is not an observed call, booking or transaction. |
Earlier MCAC observations used the label Actionability. The public framework uses Conversion-readiness to match MCAC; this naming change does not turn action information into a measured conversion.
Why the dimensions need to stay separate
An answer may cite a restaurant article without naming the restaurant in its text. That gives a visible citation with no corresponding body mention. Another answer may name the restaurant but provide its old phone number: the mention is present while the factual check fails.
Both examples are illustrative. Neither supports an estimate of how frequently a platform behaves this way.
Record the unit, denominator and uncertainty
Keep the question, full response, model, date and decision rule. For a rate, state what counts as an eligible observation and show the numerator and denominator. Failed or unverifiable observations need explicit states rather than silent replacement with zero.
Compare like-for-like questions and dimensions. This definition does not supply a validated aggregate scoring formula or establish a causal link to business revenue. See the AI visibility measurement guide for a practical reporting structure.
Sources and scope
Source-language entry updated 2026-09-22; no separate source-check date is recorded. This adaptation preserves that provenance and does not represent a new audit of the cited pages.