**Seminal Perspectives | Touch Stone Publishers | Category 546**
*Publication: Day 1 of sequence*
—
There is a thought that leaders are reluctant to say aloud, because it sounds like fear disguised as wisdom: that the thing most worth having in an organization is the thing that cannot be computed.
It is not fear. It is economics.
The National Bureau of Economic Research published its analysis of AI-powered investment funds in May 2026. The finding was specific: AI funds outperformed human-managed peers during early adoption. Then competing algorithms replicated the same strategies. The alpha evaporated. Not because the AI failed. Because the AI succeeded: and every competitor deployed the same successful AI and achieved the same result. Competitive advantage that can be computed and copied is not a moat. It is a temporary price reduction shared simultaneously with everyone else.
When the computational playing field levels: and the data show it does, consistently: what remains is what the algorithm cannot touch. The quality of the relationship between two people who trust each other. The judgment of a human who has spent years developing the contextual wisdom to address a situation the model has never seen. The willingness of a team to give their best effort for a leader who they believe sees them clearly and cares about what becomes of them.
These are not soft assets. They are the only assets that do not commoditize.
—
The mistake most organizations make is treating AI as a replacement mechanism. They deploy AI to eliminate the cost of human judgment in defined task categories, celebrate the efficiency gain, and then discover: sometimes quickly, sometimes slowly: that the human judgment they eliminated was also the judgment they needed to catch the AI’s errors, to evaluate the AI’s outputs, and to exercise the contextual wisdom that no training dataset contains.
The NBER’s companion paper on AI, productivity, and the workforce names this pattern explicitly: knowledge collapse. When an organization automates away the practice of analytical judgment, the practitioners lose the competence to audit the AI’s work. What they have built is not an efficiency system. It is a dependency. And dependencies are the opposite of moats.
The organizations that will emerge from the AI era with genuine competitive advantage are not the ones that deployed AI fastest. They are the ones that used AI deployment as an occasion to concentrate human judgment where it matters most: the relationships that sustain trust, the decisions that require contextual wisdom, the leadership conversations that determine whether people stay or go.
—
There is a leadership pattern worth naming, because it describes the failure mode that most boards and most CEOs are currently working through without understanding what they are working through.
The Declarative Board Failure Pattern works like this: the board adopts an AI ethics commitment. It assigns the topic to a committee. The committee receives positive briefings from management. The governance record reflects that AI has been discussed. The governance obligation: which is to build a reporting system that surfaces real information about real risks: is never actually met.
This is not corruption. It is the natural result of a system that rewards governance activity over governance efficacy. Declaring a commitment satisfies proxy advisor checklists. Building a functioning oversight system requires something more: a board that asks the questions management would prefer not to answer, that requires documentation rather than accepting assurances, and that treats the governance obligation as a governance obligation rather than a governance performance.
The boards that get this right are the boards that have raised their own expectations about what oversight means. They have applied the Expectation Elevation Model to themselves: not managing the governance standard, but raising it to the level that the organization requires.
—
I have watched organizations choose the efficiency narrative over the trust architecture, and I have watched what follows. It is not dramatic. It does not announce itself as a failure of leadership. It accumulates: in the workforce that disengages before anyone notices the pattern, in the client relationships that fray before the revenue impact becomes visible, in the governance record that reads as active oversight and contains none.
The organizations that address AI well are the ones led by people who understand, without requiring a regulatory crisis to demonstrate it, that the efficiency of an AI system is only worth what the human architecture around it can sustain.
The empathy is not the supplement to the strategy. The empathy is the strategy.
The full research on what this architecture requires: across the board, the CEO, the CFO, and every C-suite function: is available through Touch Stone Publishers’ AI governance research suite at the [Leadership Reinvention in the AI Era research hub](https://touchstonepublishers.com/leadership-ai-purpose/).
—
*Glenn E. Daniels II is the founder of Touch Stone Publishers Limited. He has spent three decades observing how boards, CEOs, and executive teams build: or fail to build: the leadership architectures that outlast them.*