Mario Thomas

The Board in the Machine | July 2026

The Line That Does Not Move

On 12 June, a national-security directive forced a frontier AI provider to withdraw two deployed models from every customer overnight. For the organisations that had built processes on those models, the capability did not degrade; it disappeared, with no notice period and nobody to negotiate with. It was the first observable proof of a risk that had until then only been described: that control governed elsewhere can be revised elsewhere, without consultation.

June's three articles circle a single principle from different chairs. AI and the CFO examines the first executive officer in the series, whose job has never been to produce the numbers but to stand behind them, and sorts AI's effect on the finance function into where it does the work, where it sharpens the judgement, where it operates out of sight, and where it changes almost nothing. The AI Sovereignty Trilemma returns to a framework I set out last year, now tested by events, and asks whether Boards know which of the Trilemma's costs they are paying and whether they chose it. AI and the CEO completes the executive set with the one office answerable on both counts and to the market besides, where PwC finds 74% of AI's economic value captured by just 20% of organisations, making the choice of bets, not the size of spend, the decisive act.

The pattern across the three is the distinction between agency and accountability. The doing of the work can move: to a machine, to an agent, to a provider. What cannot move is who answers when the numbers are wrong, when the capability vanishes, or when the bets fail to pay. Many organisations believed they had outsourced the operational risk of frontier AI; what they had outsourced was control, and the risk stayed home. The same logic runs through the finance function and the corner office. AI changes who produces, who builds, and who signals. It does not change who signs.

If your time is limited, I particularly recommend The AI Sovereignty Trilemma. The June withdrawal supplied, at someone else's expense, the demonstration that usually arrives only after harm, and the article sets out the questions a Board should be asking while the lesson is still free.

How is your organisation distinguishing between the work it has delegated and the accountability it cannot, and does your Board know which cost of the Trilemma it is paying?

- Mario

This Month's Reflection

A CFO's desk at night in a dark wood-panelled office, lit warm on the left by a brass desk lamp where bound financial statements lie open with a fountain pen resting on a freshly signed final page, beside a crystal glass and a leather folder, and cool blue on the right through a floor-to-ceiling window onto a screen-filled finance operations floor, with an empty leather chair between them, a visual reframe of numbers a machine produces but a human still signs

AI and the CFO: Standing Behind the Numbers the Machine Produces

Published 7 June 2026 | 12 minute read

Commitment to AI in the finance function now runs well ahead of readiness, but accountability does not wait for that gap to close. The CFO answers for the integrity of the accounts and the stewardship of capital every reporting cycle, ready or not, and AI is already inside the work that produces both. This article argues that AI changes how each of the CFO's duties is carried out, not who signs for them, sorting its effect into four groups: where it does the work, where it sharpens the judgement, where it operates out of sight, and where it changes almost nothing.

Read Article
A human hand pressed flat against a dark glass barrier at night, reaching toward an illuminated toggle switch mounted in a glass housing on the far side of the glass, out of reach, with a warm-lit city skyline glowing across water beyond, a visual reframe of a control that is visible but held elsewhere, behind a barrier the hand cannot cross

The AI Sovereignty Trilemma: When a Frontier Model Vanishes and Reality Bites

Published 14 June 2026 | 10 minute read

Every position in the AI Sovereignty Trilemma carries a cost, but only one is shown to a Board before it is paid. On 12 June a national-security directive forced a provider to withdraw two frontier models from every customer overnight, turning a structural claim into a documented event. This article argues that model availability is a continuity risk a Board must own, and that the task is not to solve the Trilemma but to know which cost the organisation is paying, and to have chosen it.

Read Article
A corner executive office in dark mahogany panelling at night, lit warm on the left by a stained-glass Tiffany lamp beside a porcelain teapot and cup, where a spread of closed leather dossiers fans across a polished desk and one is squared up in the lamplight before an empty high-backed leather chair with a fountain pen resting on it as the chosen bet, the rest angled away into shadow, and cool blue light falls from floor-to-ceiling windows onto a night city skyline and waterfront beyond, a visual reframe of the few bets a chief executive draws from many and answers for, where a machine can lay out the options but only a human can choose

AI and the CEO: Choosing the Bets That Matter

Published 21 June 2026 | 12 minute read

The market now treats visible AI adoption as proof a company is driving forward through innovation, and the chief executive is the one expected to show it. Being seen to adopt, not misleading the market, and choosing well are three demands held at once. This article argues that AI does not rewrite the chief executive's duties; it changes the conditions under which they are discharged, and the task is to choose the few bets that matter, change how the company works around them, and answer for them without delegating the accountability.

Read Article
LinkedIn X GitHub YouTube Quora Reddit Medium Pinterest Telegram RSS SubStack
Terms | Privacy | Cookies

Copyright © 2026 Mario Thomas. All rights reserved.

You are receiving this email because you signed up to receive it on mariothomas.com on {{date_subscribed}} at {{time_subscribed}}. You can unsubscribe at any time.