A recurring, self-serve survey of senior in-house AI and data leaders — Directors, VPs, and Heads of AI/Data at mid-to-large enterprises, primarily across the EU, UK, and Australia. Unlike organisational "State of AI" surveys, its unit of analysis is the leader: mandate, influence, technical currency, and access to peer learning. Respondents receive a personalised reading of where they stand relative to the field; anonymised aggregates are published as a longitudinal series.
Fifteen fixed questions per wave. The core question set does not change between waves; any amendment is versioned and disclosed, and time-series comparisons are only made within a scoring version.
Placement statistics (terciles, bands, percentiles) are computed deterministically in code from the responses of verified, consenting respondents. The interpretive narrative in each personal reading is generated from those pre-computed numbers and a fixed set of editorial materials; the generation step cannot introduce numbers of its own.
Everyone who completes the survey receives a reading. Published statistics draw only on respondents who (a) gave explicit consent for anonymised aggregate use, (b) passed verification of role seniority (work-email and role screening), and (c) provided complete responses. Duplicates are removed on email address.
Individual responses are never published, shared, or sold. Free-text answers are quoted only with separate, explicit consent, anonymised and scrubbed of identifying detail. Contact details are stored separately from analysis data and used to deliver readings and subsequent research findings; unsubscription is one click. Respondents can request deletion of their data at any time: research@longview.institute.
The Barometer is designed and run by the Longview Institute for AI Leadership. It is not sponsored by, and takes no funding from, any technology vendor. Questions about the method are welcome: research@longview.institute.
Instrument v1.0 — 15 questions.