The execution advantage is closing. The remaining moats are structural.

Work that required coordination, analysis, drafting, and reconciliation now runs at the speed of a model query — across every company, simultaneously. What is commoditized cannot be a moat.

What is being commoditized

Execution — the capacity to produce output quickly and at scale — is no longer scarce. Models produce drafts, summaries, analyses, coordination, and routine decisions at marginal cost. Organizations that compete on execution are competing for a prize that is disappearing.

This is not a productivity shift. It is a relocation of where value is created.

What is becoming scarce

Four things are appreciating under AI pressure. Each one defines a different kind of advantage. Each one requires deliberate design.

Judgment.

The capacity to commit to an irreversible decision under incomplete information. Reversible work moves to the model. Commitment does not.

System architecture.

The design of how decisions, information, capital, and authority flow inside the organization. Architecture determines which AI leverage converts into a structural advantage and which dissipates into cost savings.

Decision topology.

The map of who owns what. AI exposes every point in an organization where authority is unclear and punishes the ambiguity with latency. Clean topology is the difference between faster output and structural speed.

Reputational coherence.

In a market where information moves faster than explanations, a consistent signal between what a company does, says, and decides becomes a financing condition. Incoherence is now priced.

The gap this newsletter will hold

Most executive teams already describe themselves as AI-native. Allocation — of capital, of authority, of attention — rarely confirms the description.

The gap between description and allocation is the subject of this newsletter.

Fifty-two weeks, arranged in four movements:

Relevance.

Why most AI initiatives fail at the business-model level before they fail at the technology level.

Capital.

How valuation, funding, and M&A are being quietly rewritten around AI leverage.

Reputation.

Why coherence is becoming infrastructure — and silence, a decision.

Growth.

What separates organizations that compound from organizations that merely scale.

What this newsletter is not

Not a briefing on tools, models, or prompts. Not a productivity resource. Not commentary on industry news. These are not scarce goods. The internet produces them at commodity rates.

This publication exists to do one thing: name the structural shifts precisely enough that boards, founders, and investors can act on them before the shifts become consensus.

The verdict that will close every issue

Preserve / Decommission / Re-author — applied to whatever is under examination that week. A role, a pricing model, a policy, a governance structure, a product line, an assumption.

The one question

For the next fifty-two weeks, this newsletter holds every structural decision against one question:

Is this choice preserving an advantage that still compounds, or preserving a habit that no longer does?

Most organizations answer quickly. The answer is usually wrong. The distinction matters only to organizations willing to stay with the question long enough for allocation to catch up with description.

OM on Architecture is read weekly by founders, boards, and investors building companies that compound.