Where the data comes from
Every InsightsOut workshop asks a short set of optional questions at registration, and increasingly a three-question pulse afterwards. Each event is one wave in an ongoing study. As of the current dataset (version …): … waves, … registrations, … unique people, ….
What we ask
From August 2026, every event carries the same five core questions so waves are comparable over time: who you are right now; the biggest change moving through your work; the hardest part of it; the emotions most present; and what you want to walk out with. One or two rotating questions per event test a specific hypothesis. Earlier events used different forms; where we count across them, we map answers to a common set of themes and say so.
How we code it
Answers are coded to a stable set of themes — decision uncertainty, structural overwhelm, role and identity, leading others through change, peer processing, AI adoption, founder existential pressure, positive and negative emotion, and conflict. Repeat attendees are tracked, not double-counted. Codes are extended, never renamed, so counts stay comparable.
How we cross-check it
Each week we read what credible institutions publish on AI and work — Gallup, Microsoft, MIT Sloan, Stanford HAI, McKinsey, BCG, Deloitte, Pew, WEF, OECD, SHRM, Anthropic, and others. Every finding we publish is paired with at least one external source and states the relationship plainly: confirms, complicates, or contradicts. The pieces we value most are the last two.
What we will and won't publish
- Aggregates and anonymized quotes only. No names, companies, or identifying situations. Nothing from coaching sessions or client work ever enters the research dataset.
- A percentage appears only when the segment has at least 20 people, and a headline claim only at 50 or more. Smaller signals are labeled directional.
- Refuted hypotheses are published, not deleted.
Limits
…
Living papers and versions
Our papers are versioned like preprints. A patch (v1.0 → v1.1) means a new wave moved the numbers but not the claims. A minor edition (v1.1 → v1.2) means a hypothesis changed status or a section was added. A major edition means the questions or the thesis changed; the old paper stays up, marked superseded. Every paper shows its version, sample, and a "what changed" log; PDFs carry the version in the filename and older editions stay downloadable.
Cite or reuse
Cite as: Imani, N. (year). Title (version). InsightsOut. URL. Reuse of aggregate figures with attribution is welcome. For the underlying dataset or collaboration, get in touch.