AACF

AI Adoption Controls Framework

What an organisation needs in place to adopt AI well, who owns it, how far along it is, and how that lines up with the standards it answers to: 93 controls in 11 domains.

Take the self-assessment Browse the controls

What it is

A controls matrix for AI adoption: the organisational layer that sits above technical and security control sets. It covers strategy, governance, risk and ethics, the portfolio of uses, the operating model, data, technology, security, people, culture and measurement.

Each control has an objective, five level descriptors from Initial to Optimising, with Level 3 as the foundation every organisation should have, and Levels 4 and 5 for going beyond it, assessment questions, the evidence an assessor would expect, an accountable owner, dependencies on other controls, short guidance on reaching Level 3, and cross-references to published standards.

How it is used
  1. Assess. Score each control from 0 to 5 against its level descriptors, in the browser or in the assessment workbook.
  2. See the gaps. Mean level and the share of controls at target, by domain.
  3. Plan. The controls below target, in dependency order, become the improvement roadmap.
  4. Show alignment. Each control is cross-referenced to the standards below, so one assessment speaks to several of them.
Status

Version 1.0.0. Open for comment: tell me what is wrong or missing.