The Art of Defensible Data and AI Governance - Managers

Most of the governance work in your practice has already been done. You just cannot prove it.

That gap is where this Thursday starts.

Howard has spent the past few months building a full-day workshop for people running a data or AI practice: how to assess your ethical maturity honestly, and what to do with the policy that comes out of it. Asking a practice head to give up a whole day is a hard sell, so he have split the material across four sessions on consecutive Thursdays.

Session 1 makes the case for why any of this matters, and it is the part people push back on hardest.

Take two organisations that behave identically. Only one of them can show it, and only that one has a defensible position.

The difference between them is rarely effort. It is filing.

This Thursday at 16:00 SAST Howard will cover:
● What a reviewer sees when they assess you, and why "we basically do this" scores as nothing

● The four tests an artefact must pass before it counts as evidence: dated, attributed, owned, and locatable by someone who has never seen your systems

● The three ways practices lose the argument. Absence is the rare one. The unevidenced claim is far more common, and it reads as maturity from the inside while reading as negligence from the outside

● What King V asks of you when technology is a strategic pillar, including the half almost nobody reports upward: value enabled, and the opportunity you chose not to take

● Why proportionality and an honest "not applicable" are legitimate governance positions and not shortcuts

The exercise: bring one thing your practice "basically does" and we will test it live against the four tests. It is uncomfortable. It is also the quickest honest look at your own baseline you will get this year.

One rule runs through the whole series.

The instrument informs the judgement; it never makes it. No score in this framework closes a gap, approves a project or certifies anyone. A named person does that, and their name goes on the record. If you want a maturity score to drop into a board pack, this series will probably frustrate you.