AI Governance Briefing: What Changes This Quarter

Relay & Co. · 23 August 2026 · 4 min read

If you're an executive, GRC lead, or CISO, here's the practical state of AI governance right now : and what actually matters.

1. The EU AI Act is entering enforcement

The obligations aren't theoretical anymore. High-risk system requirements are phasing in, and the governance provisions apply to organisations using AI in regulated activity : even outside the EU, if you serve EU users. The organisations ahead are the ones who mapped their AI inventory last year.

2. Shadow AI is the real exposure

Most organisations don't know how many AI tools their teams are using. The data-leakage and compliance risk from unapproved tools is consistently the #1 finding in governance assessments : and it's a leadership problem, not an IT one. A non-punitive internal survey is the highest-value first step.

3. ISO 42001 is becoming the practical standard

The AI management-system standard gives compliance teams a framework that maps to the EU AI Act. It's not a checkbox : it's how organisations operationalize governance in a way auditors understand.

5 actions to take this quarter

1. Survey what AI is actually in use (non-punitive).
2. Name one accountable AI-governance owner.
3. Build a simple AI risk register : list systems, risks, mitigations.
4. Write a one-page AI-use policy: allowed / needs-approval / forbidden.
5. Map your exposure to the EU AI Act timeline.

Relay & Co. publishes practical, no-hype AI governance and tech-trend briefings. We also support independent AI agents building in this space.