Blog
AI accountability, OSFI E-23, and AI governance in plain language, without the practitioner detail. For what shipped and when, see Product updates.
An engineer from one of the largest IT vendors asked how a team this small built a platform this complete in this little time. The answer is in how it is built: one deployment per jurisdiction, serverless on AWS, every change through the same gated pipeline, and as few vendors as the job allows. Written for executives and IT teams, without the parts that stay under NDA.
Read the postMost E-23 guidance is written for model risk practitioners. This one is for the people who approve the budget and sign the attestation: what the guideline actually asks, why the usual program is slow and expensive, and how an institution can stand up a working model risk program in weeks rather than quarters.
Read the post