CHART
An AI governance instrument I authored. Every mandate I run carries it.
CHART assesses whether an organisation can evidence meaningful human oversight of its AI systems. That's the obligation Article 14 of the EU AI Act places on every high-risk deployment, landing in December 2027. The deadline was deferred for a reason: the tooling to evidence oversight doesn't yet exist. CHART is built to close that gap, applying measurement rigour from my psychology research to governance claims that are usually taken on trust.
The framework is openly published: the five dimensions are set out in a free 15-question self-assessment, and the research behind it is public as an SSRN preprint. The assessment protocol, scoring and calibration are the professional engagement. Inside an interim or fractional mandate, CHART means AI adoption decisions get made against a tested oversight instrument rather than a vendor's checklist, and the board gets an evidenced position it can stand behind.
Commercial engagements run through beò, the research company I founded.
If this is your problem
If your organisation deploys AI in anything the EU AI Act would call high-risk, and you couldn't evidence your human oversight to a regulator tomorrow, that's the conversation to have. Start with the free self-assessment: fifteen questions, five dimensions, an honest read on whether you should be worried.