Why Executives Must Treat Intelligence As An Asset

AI has quietly crossed a threshold. In many African organisations today, AI is no longer an experiment or a pilot. It is embedded in underwriting, credit decisions, claims processing, pricing, forecasting, and operational control. Shaping outcomes repeatedly, improving through feedback, and delivering real value. Economically, it could be said that AI is behaving like capital. Yet in financial statements, it is still treated mostly as a cost. This gap, between how AI behaves inside the firm and how it appears in reports, has become one of the most misunderstood tensions in modern enterprise management.

According to Michael Compton’s newly published white paper, AI as Governed Capital, the problem is not that accounting is broken. The problem is that governance has not kept pace. “The constraint on AI capitalisation is not accounting. It is governance.”

What, then, is “governed capital?”

Governed capital refers to assets whose value is preserved, extended, and trusted because they are deliberately managed, controlled, and evidenced over time. Traditional capital, such as plants, equipment, and financial instruments, earns recognition because ownership is clear, performance is measurable, and deterioration is monitored. AI can meet these same expectations, but only when it is governed intentionally.

In the paper, AI qualifies as governed capital when it demonstrates:

• Identifiability: discrete models, systems, or intelligence units that can be named and tracked
• Control: clear ownership of data, models, and derivative outputs
• Versioned lineage: documented evolution, updates, and decision logic
• Attributable contribution: evidence linking AI performance to financial or operational outcomes

Without these, AI remains an expense. As governed capital, AI becomes intelligence that compounds.

Accounting Conservatism Is Not The Enemy

An interesting key insight of the paper is its defence of conservative accounting, especially in African markets. Accounting standards insist on evidence, not on enthusiasm. They are designed to prevent speculative valuation and protect trust. However, AI often fails recognition tests not because it lacks value, but because its development is fragmented, undocumented, or dependent on vendors. Curious metrics these. Compton writes, “Accounting’s conservatism is a feature, not a bug.” The strategic mistake many organisations make is waiting for accounting standards to change. The smarter move is to supply the evidence that those standards already require.

The Hidden Cost Of Ungoverned AI

When AI is not governed as capital, risks crop up. AI systems cannot be audited or transferred cleanly. At the same time, institutional knowledge walks out when staff or vendors leave with no one to track it. Any performance gains cannot be defended during due diligence, and organisations lease intelligence instead of owning it. This matters profoundly on the continent where digital sovereignty, regulatory scrutiny, and investor confidence are tightly linked. AI acquired “as-a-Service” may deliver functionality, but it does not build institutional value.

The AI arms race would have us believe Africa is at a disadvantage. Not according to this report. Because many African AI systems are newer, they are less tangled in legacy tech. This creates an opportunity to design governance properly from the start with registries, lineage, stewardship, and performance tracking built in, not bolted on. “African institutions may reach capital-ready evidence sooner than some Organisation for Economic Co-operation and Development (OECD) firms,” Compton notes, “precisely because their deployments are newer and more governable.”

The average CIO’s investment incorporates AI as a business project. But one of the most practical recommendations in the paper is the idea of treating AI systems as units of account on the ledger. This does not replace financial reporting. If anything, it complements it. Each AI system is assigned a steward, a defined purpose, a useful life, maintenance and retraining expectations, and performance and degradation indicators.

This ultimately matters to CEOs and CFOs because, beyond the technology function, the message is clear: AI strategy is now capital strategy. As Compton states plainly, “Governance, not scale, is what makes AI durable, portable, and auditable.”