Applied Erudition™

Evidence That Survives a Changed Assumption

A finished deliverable no longer tells you much about the person who submitted it. AI did not create that problem; it removed the last excuse for ignoring it.

Applied Erudition is built on a different unit of evidence: work a person has to stand behind after the conditions change.

The case is parameterized. Each candidate draws different numbers from the same published structure, so a circulated answer is wrong rather than useful. The candidate builds a working model and commits to a recommendation in writing. AI tools are allowed; the call is theirs.

Then the work can be called to defense. After submission, the candidate receives a question they could not have prepared for and answers it in their own voice: an assumption moves, an objection arrives, a number changes. The work and the defense are scored together against a published rubric, never by the person who coached the work. AI assists the scoring; it does not decide, and student work never trains a model.

The defense is not an integrity check bolted onto an assignment. It is the evidence. A good model shows the work was done. Holding, revising, or abandoning a recommendation when the ground shifts shows whether the person can be trusted with the next one.

We are deliberate about what this supports. A defended performance under stated conditions supports a claim about that performance. It does not, on its own, support claims about job readiness or long-run outcomes, and we will not make those claims before the evidence exists.

Applied Erudition begins with business, and it keeps scoring apart from coaching.

We write about building it, issue by issue, on Substack.

Follow the work

We're building Applied Erudition one domain at a time. If you want to follow along, as an institution, a coach, an employer, or someone who would rather have their capability measured than assumed, leave your email. We'll write when there is something real to report, and you'll hear first when Applied Erudition opens in January 2027. That's the whole promise. Reply to any message to come off the list.

I'm here as

See our privacy policy for how we handle your email.