nikoalho/ai-recommendation-index

AI Recommendation Index

An open quarterly dataset tracking which software brands AI assistants mention and recommend first across fixed buying prompts.

Q2 2026

• —144 complete responses
• —6 AI models
• —24 fixed prompts
• —4 software categories
• —English and Finnish

Each record includes the complete prompt and response, collection timestamp, model identifiers, normalized brand mentions, first explicit recommendation, and returned citations.

Use cases

• —AI recommendation and brand-visibility research
• —Generative engine optimization measurement
• —Cross-model comparison
• —Citation-source analysis
• —Software-category benchmark studies

Files

• —responses.csv: response-level table used by the Dataset Viewer.
• —responses.json: nested response-level export with release metadata.
• —summary.json: aggregate mention and first-pick counts by category.
• —manifest.json: release, license, provenance, and distribution metadata.
• —DATA_DICTIONARY.md: field definitions and interpretation notes.

Limitations

The data is a dated observation of non-deterministic model behavior. It is not a timeless vendor ranking. Model versions and live-search results change, and later waves should be treated as new measurements rather than exact reproductions.

Methodology

Read the complete methodology, category definitions, and interpretation guidance.

License and citation

Licensed under CC BY 4.0.

Alho, N. (2026). AI Recommendation Index: Q2 2026 (Version 2026-q2) [Data set]. Zenodo. https://doi.org/10.5281/zenodo.21420271

Preferred attribution:

AI Recommendation Index by Niko Alho — https://nikoalho.fi/writing/ai-recommendation-index/

Source repository: https://github.com/alhoniko/ai-recommendation-index

DOI: https://doi.org/10.5281/zenodo.21420271