What it is.
Manoj Palanikumar built this around one finding: blended AI visibility numbers hide 3x-10x gaps between platforms. Step one is a matrix of citation rate by platform and by query category, carrying a time estimate of 4-6 hours and a named owner.
Step two scores content against three signals, so answer-first formatting where the core claim lands in the first 100 words, named authorship, and data a model can lift. Step three is the technical pass over schema and platform coverage.
TripleDart's own guide, and TripleDart sells the audit as a service. Every scoring criterion is published, so you can run it yourself.
What you get.
- A platform-by-category visibility matrix as the first output.
- A content scorecard against three named signals, with the pass and fail pattern written out for each.
- A schema audit checklist with the common failures listed.
- Time estimates and a team assignment per step, starting at 4-6 hours for the baseline.
- A worked client example showing a 10x spread between the best and worst platform.
How to set it up.
Pick your query categories before you query anything, because the matrix only helps if the categories match how buyers ask.
Run the same query set through each AI platform and record citations per platform per category.
Read the spread, since a healthy aggregate can hide a platform where you never appear at all.
Score each key page on answer-first formatting: does the core claim sit in the first 100 words with no preamble.
Check named authorship and whether the page carries data a model can quote.
Audit schema last, and validate every type the page claims.
Use cases
See the gaps a blended score hides
Build the platform-by-category matrix first and read where you are missing.
Score the content itself
Run the scorecard against the named signals for each page you care about.
Choose where to spend the quarter
Use the matrix to pick the category with the widest gap.
Best for
A first AI visibility audit
You have a blended number from a tool and no idea which platform or query type is weak.
Reporting to a skeptical exec
A matrix with a 10x spread across platforms argues better than a single score does.
Auditing a competitor
All three steps run against somebody else's pages with no access to their analytics.