The Triple D is short for the Drift versus Design Diagnostic. It takes the argument in SuperSkills that organisations either design their relationship with AI or drift into it, and asks the same question at the scale of one working week. The argument it comes from is set out on drift versus design, and the matrix is the version of it for teams.
It asks how much of your working life you are choosing and how much of it is running on defaults. Ninety-four questions in nine parts return a score out of a hundred, a band running from Deep Drift to Deep Design, one of nine archetypes, and the one of the seven SuperSkills where your answers showed the widest gap between where you are and where you could be. It is a structured prompt to think and not a measurement of a person. Nothing published about it reports a reliability coefficient, a validation sample or a test-retest figure, and it is not trying to be a psychometric instrument.
About twelve minutes. Ninety-four questions in nine parts, no account, no payment and no time limit. Nothing is saved in your browser between visits, so leaving the page means starting again.
Your answers are scored in your browser and the ninety-four of them are not transmitted anywhere. What is sent, when the results appear, is a summary: the score, the band, the archetype, the weakest skill and the answers to the handful of questions that carry no score, which are there to see how people read the instrument itself. If you give a first name and an email address at the end so the result can be sent to you, those are sent with the summary. It goes to one spreadsheet. There is no advertising pixel on this page, and the analytics that is here counts actions and not people.
Rahim writes a free weekly letter on AI and human capability, read by 25,000 people every week since 2017. It is called Box of Amazing, it is free, and it is one click to stop.
First published 2026-09-27. Written and checked by Rahim Hirji, author of SuperSkills (Kogan Page, 2026). The Triple D is a structured prompt to think about your own defaults. It is not a psychometric instrument and does not report a validated score.