The AI quietly crossed the threshold. Today, in many African organizations, AI is no longer an experiment or a pilot. It is built into underwriting, credit decisions, claims processing, pricing, forecasting, and operational management. Iteratively shape results, improve through feedback, and deliver real value. Economically, you could say that AI is behaving like capital. However, most of it is still treated as an expense in financial statements. This gap between how AI operates within an enterprise and how it appears in reports has become one of the most misunderstood tensions in modern enterprise management.
According to Michael Compton’s new white paper, AI as Governance Capital, the problem is not that accounting is broken. The problem is that governance has not kept up. “The constraint on the use of AI is not accounting; it is governance.”
So what is “government capital”?
Governance capital refers to assets that are intentionally managed, controlled, and evidenced over time so that their value is maintained, extended, and trusted. Traditional capital such as plant, equipment, and financial instruments are valued because ownership is clear, performance is measurable, and deterioration is monitored. AI can meet these same expectations, but only if it is intentionally controlled.
In the paper, AI qualifies as governance capital if it demonstrates the following:
Identifiability: Discrete models, systems, or intelligence units that can be named and tracked Control: Clear ownership of data, models, and derived outputs Versioned lineage: Documented evolution, updates, and decision logic Attributable contribution: Evidence linking AI performance to financial or operational outcomes
Without these, AI will still be expensive. As governed capital, AI becomes a complex intelligence.
Accounting conservatism is not the enemy
An interesting and important insight of this paper is its defense of conservative accounting, especially in African markets. Accounting standards value evidence, not enthusiasm. These are designed to prevent speculative valuations and protect trust. But AI often fails recognition tests not because it has no value, but because its development is fragmented, undocumented, or vendor-dependent. It’s an interesting metric. Compton writes that “accounting conservatism is a feature, not a bug.” A strategic mistake many organizations make is waiting for accounting standards to change. A smarter move would be to provide evidence that those standards already require.
The hidden costs of ungoverned AI
Risks arise when AI is not governed as capital. AI systems cannot be cleanly audited or transferred. At the same time, when staff or vendors leave, organizational knowledge is lost with no one to keep track of. Performance improvements cannot be defended during due diligence, and organizations lease intelligence rather than own it. This is critical for a continent where digital sovereignty, regulatory oversight and investor confidence are closely intertwined. AI acquired “as a service” may provide functionality but will not build organizational value.
One would think that Africa would be at a disadvantage in the AI arms race. Not so, according to this report. Many of Africa’s AI systems are newer and therefore less intertwined with legacy technologies. This creates an opportunity to properly design governance from the beginning, with registry, lineage, stewardship, and performance tracking built-in rather than added-on. “African institutions may reach evidence of capital reserves sooner than some organizations in the Organization for Economic Co-operation and Development (OECD),” Compton points out.
The average CIO’s investment includes AI as a business project. But one of the paper’s most practical recommendations is the idea of treating AI systems as units of accounting on a ledger. It is not a substitute for financial reporting. If anything, it complements it. Each AI system is assigned a manager, a defined purpose, a useful life, maintenance and retraining expectations, and performance and degradation metrics.
This ultimately matters to CEOs and CFOs. Because beyond the capabilities of the technology, the message is clear: an AI strategy is now a capital strategy. As Compton articulates, “Governance, not scale, makes AI durable, portable, and auditable.”
