Evidence Over Attestation, Part Four: Where the Real Moat Lives

The realization that matters most in the Intelligence Era is who becomes the builder. In the previous era, solving a critical problem required translating human domain knowledge through layers of non-domain engineers, project managers and process owners, and the customer’s need was diluted at every layer.

Today the barriers to technical implementation are collapsing rapidly. The nurse who understands why the seven o’clock handoff fails. The agricultural manager who knows which soil-moisture reading is ignored. The compliance officer who spots regulatory drift in a complex contract. The warehouse supervisor who sees where the packaging line disintegrates. These domain experts are the real founders of the next decade, and the people a CEO should be looking for.

They no longer need to wait for a multi-year capital allocation. A set of agents can provide the architecture, the data design and the plumbing in days rather than quarters. But the human must bring the standard of proof: the judgment to interrogate the synthetic artifact, the character to reject the dishonest metric, and the craft to protect the vulnerable person on the other side of the screen.

The disciplines in Part Two are a syllabus, and a syllabus can be examined. Ask a candidate, or a partner, or a member of your own leadership team, to walk through the last time a system they were responsible for reported success and was wrong. Listen for whether they found it or a customer did. Listen for the question they asked, and whether it was theirs or the dashboard’s. Listen for what they told the customer, and when.

Ask what “done” meant on their last significant piece of work, and whether that definition existed before the work started. Ask for the artifact. The people you want will have it, will be slightly embarrassed by how plain it is, and will be able to explain what it proved from the customer’s side.

Ask what they refused to build last quarter, and why. Drucker’s test of an effective executive was never how much they did. It was whether they did the right things, and could say which things they chose not to do so that the right ones got done.

And ask what they think of agents. The wrong answer is fear, and the other wrong answer is faith. The right answer is that the agents are extraordinary at everything except knowing what matters, and that knowing what matters is the job they intend to keep.

• Audit your attestation gap. Inspect your internal AI projects. Are teams measuring agent success by generated volume, such as lines of code, automated summaries and closed tickets, or by verified outcomes for customers?
• Ban self-certifying workflows. Mandate that no autonomous agent, AI pipeline, or team may evaluate its own output without an external, independent verification gate stated in the customer’s terms.
• Empower the domain experts. Move generative AI budgets away from generic, centralized sandboxes and into the hands of the people who know the customer’s problem from the inside.
• Reward the honest zero. Promote the executives and teams who halt a deployment to fix a measurement flaw, rather than those who ship a flattering estimate.

Every domain has a problem like the ones in this essay, and every domain has a person who understands it better than any team that could be hired. For most of the history of software, that person waited for a budget that never came, or watched their need diluted through layers of translation until what shipped solved something else.

That wait is over, on one condition. Agents can increasingly supply implementation speed and technical breadth. They cannot supply the standard by which the result deserves trust. The person who knows the problem has to bring that judgment. The agents will not bring the standard. They will not bring the refusal to believe the obvious story. They will not bring the decision to hold a release because the instrument is not honest yet, or to publish a zero rather than a guess, or to tell a customer the truth before the fix is ready.

This is not speculation. It is the record of one Tuesday in September 2026, fifteen products in production, and three years of the same rule. Those are the skills. They are learnable, they are inspectable, and they are the whole of the moat.