# Australian Professional Services AI Visibility Study: public research data

Study version: `2026-08-29-v1`

This package contains the frozen prompts, method, codebook, privacy-safe observation-level coding and independently recalculated aggregate results used in Integral Media's published study.

## Public-data boundary

Named firms, full answer text, raw captures, URLs, domains and reviewer notes are excluded. This prevents the dataset from republishing unverified claims or personal information while retaining the observation design, denominators and coded result fields needed to inspect the findings. Counts show how many visible links or source classes were coded; they do not identify or endorse the sources.

## Interpretation

This is a dated observational snapshot of consumer-facing interfaces, not a traffic estimate, permanent ranking study or assessment of firm quality. Results can vary by account, location, session, product mode and time. Integral Media sells SEO and AI-visibility services to organisations in this sector; that commercial interest should be considered when interpreting the work.

## Files

- `prompts.csv`: frozen version-one prompt set.
- `methodology.md`: collection, review, verification, limitations and disclosure.
- `codebook.md`: definitions for the underlying coding fields.
- `observations.csv`: privacy-safe observation-level coded register.
- `summary.json`: recalculated aggregate denominators and results.

The build and validation utility is versioned in the website repository at `scripts/build-public-ai-visibility-data.mjs`.