What the Epistemic Architecture Means for Coach Development
The design constraint that separates Orin from every other AI coaching tool: nothing enters a development record until a coach accepts it. Here's why that matters.
Epistemic architecture refers to the design of how knowledge is generated, validated, and recorded in a system. For AI tools in coaching contexts, this isn't an abstract concern. It has direct implications for professional ethics, coach development, and what the resulting records actually represent.
Orin's core architectural commitment is this: nothing generated by AI enters a coach's development record until the coach has reviewed it and decided to accept, edit, or discard it. This sounds simple. In practice, it's a constraint that shapes everything about how the system works.
Why this matters for professional ethics
Coaching relationships are built on confidentiality and trust. When AI processes a coaching session, it generates observations — candidate assessments of what it detected in the session. These observations are hypotheses, not facts. They're based on pattern matching against text.
The coach who was in the room has information the AI doesn't: the energy in the conversation, what the client was risking by saying what they said, what context explains why this moment was different from the transcript of a similar moment in a different session.
If AI-generated observations flow automatically into records without coach review, two things happen: the record contains AI judgments as if they were professional assessments, and the coach is disempowered from the development process that the record is supposed to support.
Why this matters for coach development
The goal of a development record is to support coach growth, not just document it. When coaches review AI observations and decide what to accept, they're engaging in professional reflection. The act of evaluating an AI observation — agreeing, disagreeing, modifying — is itself developmentally valuable.
A system that bypasses this process in the name of efficiency is trading away the most important part of what the record-keeping is for.
The design implications
The epistemic architecture isn't just a feature. It's a constraint on what other features are possible. It means Orin cannot offer certain kinds of automation that competitors offer — bulk session processing without review, automatic progress tracking updates, AI-generated supervisor reports that go directly to program administrators.
These are design choices, not limitations. Every one of them reflects the same underlying commitment: the coach's judgment is what the system is trying to support, not what the system is trying to replace.
In a field where "AI coaching" tools are multiplying fast, being specific about these commitments matters. Not all AI tools are the same. The architecture is the point.