Nick Tchan: Expertise by design

Real expertise involves judgement, pattern recognition and decisions made under ambiguity, and it usually takes years of apprenticeship. Nick wanted to shorten that journey: “Instead of treating expertise as a mysterious process, I wanted to treat it as something we can design.”

Nick built an agentic learning loop that draws out an expert’s tacit and explicit knowledge and turns it into structured learning. This involved several agents:

• The Questioner keeps the focus on a single whole task.
• The Analyst reads documents for contradictions, hidden rules and unspoken assumptions.
• The Interviewer draws out the judgements, decision trees and mental models experts use but rarely put into words.
• The Scenario Builder turns those into decision landscapes, with several pathways, plausible distractors and realistic ambiguity.

He built them with Supernaut, Foundations’ AI prompt engineer, using his own approach, including how he interviews experts and what they notice that others miss.