The Director's Lens.
A director's perspective on AI, technology, cyber security, capital, and the boardroom.
Each edition takes a real-world governance moment — a regulatory shift, a published filing, a board decision — and reads it through a director's lens. Five structured sections, one personal conclusion.
Researched and drafted by Brad's agentic AI team. Edited and published by Brad Ferris.
The ACTU has called for Australia's regulator to vet AI models before deployment. Whether or not the proposal lands, the political ground under AI deployment just shifted. Boards should treat the piece as a leading indicator and act now.
74% of companies plan to deploy agentic AI within two years. Only 21% have a mature governance model for autonomous agents. That gap is where organisations will be exposed — and where directors need to act before, not after, deployment.
Ninety-five percent of organisations have an AI strategy. Only 8% have measurable ROI. That 87-point gap isn't an AI problem — it's a governance problem, and it's showing up on directors' watches.
McKinsey's 2026 AI Trust survey reports trust maturity is rising on average, but only one-third of organisations hit a meaningful score on governance and agentic-AI controls. The blind spot in their analysis is more revealing than the findings: governance is treated as a management function. Directors are absent. That mirrors how most organisations are actually structuring this work — and it is exactly the gap a board has to close before the next agentic deployment is approved.
Australian companies are investing heavily in AI and worker adoption is rising — but enterprise productivity gains remain elusive. The problem isn't the technology. It's that boards are approving AI spend without overseeing the organisational transformation that determines whether that investment ever returns value.
The NDIS ran from a $14 billion estimate to $52 billion without a meaningful course correction. That is not bad luck — it is a governance system that failed to apply constraint when an institution drifted from its stated purpose. Every board in Australia should be asking the same question of its own capital commitments.
Telstra's AI board agent solves a genuine governance problem — information overload. But when every director prepares using the same AI, on the same data, with the same framing, a quieter governance problem appears in its place.
Anthropic's Mythos breach isn't really an Anthropic story. It's a stress test of third-party AI governance — and most boards are failing it without knowing it.
Most board-level AI guidance assumes an enterprise risk function and a dedicated CISO. For SME directors the real governance exposure is closer and more urgent.