Brandilite research / 2026

The 2026 AI Product Search Evidence Index.

A transparent, fixed-cohort view of the domain-level search evidence surrounding 100 AI products. Inspect the records, compare the 78 eligible scores, and see exactly why 22 scores are withheld.

Proxy v0.1 / measured inputs

One score, three evidence pillars.

The score normalises three groups of Semrush US domain signals within the eligible cohort. A rank is shown only when the record can support a like-for-like domain comparison.

40% of proxy score

Google AI Overview footprint

Normalised evidence of keywords associated with Google AI Overview results in the dated database snapshot.

40% of proxy score

Commercial and organic evidence

Commercial positions and traffic plus organic keyword and traffic evidence for the associated domain.

20% of proxy score

Authority and referring domains

Domain authority and referring-domain evidence, normalised against the eligible cohort.

Not measured by proxy v0.1

Direct assistant visibility

No ChatGPT, Claude, Perplexity, Gemini or Copilot answer observations are used in this version.

How to read a record

A withheld score is part of the evidence.

The directory keeps all 100 records visible. It does not turn missing, broad or non-comparable domain evidence into a zero.

78

Comparable domain evidence

The record receives a proxy score and a rank among the 78 eligible records. Expand its receipt to inspect the measured inputs and strongest evidence pillar.

View comparable records
22

Score withheld

The product remains in the cohort with a reason, such as a broad parent domain, multiple products sharing one domain, insufficient product specificity or unavailable required evidence.

View withheld records

Cite the dataset with its limits intact.

Brandilite. The 2026 AI Product Search Evidence Index. AI Search Evidence Proxy v0.1; Semrush US snapshot, 26 August 2026; 100 products, 78 comparable ranks and 22 scores withheld.

Find your product or request a review.

Search the full evidence set first. If the product is missing, submit it for an eligibility and evidence review. Submission is free and cannot purchase placement or rank.