Curated, not exhaustive. Source-attributed. We show our working.
Each theme links to a citation list of the most defensible sources we've drawn on. The visitor leaves knowing Attuniq is grounded in real research, not vibes.
The qualitative literature on rater cognition — differential salience, criterion uncertainty, information integration — that informs our rubric design.
Sources · BMC Medical Education · PMC qualitative studies
The research on the gap between trained and demonstrated capability — across healthcare, vocational training and corporate L&D.
Sources · CIPD · LinkedIn Learning · Tandfonline
Public reporting and field submissions on the gap between hours funded and outcomes evidenced — including ACTA's work on the EAL Framework and AMEP teachers' voices.
Sources · ABC News · ACTA submissions · Home Affairs
Cognitive load, fight-or-flight response, and interview performance — and the evidence that practice under pressure is what changes the pattern.
Sources · Adria Solutions · Quora forum data · social-anxiety research
QA scoring consistency, calibration, and the gap between training and live-call performance.
Sources · Balto · NiCE · Sharpen CX
The harder discipline is publishing evidence about how the platform performs.
What was reviewed, what changed, what didn't.
For the persona library — what was reviewed, what was retired, what was added.
From pilot partners, with consent. Anchored in rubric-aligned scoring, not promotional language.
WCAG compliance scope, tested where, by whom, what was found.
What's in the library, when each persona was added, when it was last reviewed.
Cadence: at least annually, ideally per validation cycle.
"The day we describe ourselves as 'unbiased AI' is the day to stop trusting us."
Read our principles →