Original methods, technical experiments, datasets, and AI search observations.
4 published resources
Original data: AI answer engines have near-fixed 'default' picks per category, but a volatile middle — and household names like Salesforce and LastPass show up far less than you'd expect.
This research note defines the observation pipeline, parsing rules, and limitations behind MentionPop weekly GEO and Studio reporting.
We tested n=4 AI crawlers against live robots.txt and public crawler-control files. All were allowed with HTTP 200—technical access only, not indexing or recommendation proof.
One ChatGPT answer is a data point, not a KPI. Variability is structural—your measurement design should account for it.
MentionPop field notes
New research, implementation notes, and practical GEO/AEO methods. Unsubscribe whenever you want.