TRAININGPlatform layer

MENTOR

The training layer. Mastery Engine for Navigated Teaching, Objectives, and Reinforcement: an adaptive learning agent that turns any source material into a structured, level-appropriate learning experience, and proves understanding by explanation, not completion checkmarks.

Design principle

Understanding is proven by explanation.

MENTOR's guiding standard comes from Richard Feynman: if you can't explain it simply, you don't understand it well enough. Every session is built around that test.

Two rules set MENTOR apart: gaps are always attributed to the material or the explanation, never to the learner, and the agent operates as a curious thinking partner, not an evaluator.
Phase 01

Concept analysis

Ingests pasted text, uploaded documents, or URLs and extracts 5 to 10 core concepts in dependency order, with plain-language and technical summaries, real-world analogies, and common misconceptions for each.

Phase 02

Learner calibration

Classifies each learner as novice, intermediate, or advanced through self-declaration confirmed by targeted probing, rather than taking their word at face value.

Phase 03

The Feynman Cycle

Explain simply, have the learner teach it back, diagnose the gaps, and re-explain from a new angle until understanding is proven.

Phase 04

Mastery gates

No advancement until the learner can explain the concept in their own plain language. Completion is earned, not clicked.

Phase 05

Continuous progress tracking

Follow-on questions are treated as signals, classified as clarification, application, edge case, or challenge, with pacing and level adjusted dynamically.

Phase 06

Mastery summary

Every session closes with each concept restated in one plain sentence, strengths highlighted, gaps flagged for follow-up, and an optional study note written in the learner's own language patterns.

Design

Never blames the learner

Gaps are attributed to the material or the explanation, keeping learners engaged instead of defensive.

Design

Domain-agnostic

Works equally well for cybersecurity, compliance, or general enterprise content. Whatever the source material, the method holds.

Design

Single-model architecture

The full learning context stays in one window, improving gap detection and progress tracking accuracy over multi-model orchestration.

In the platform

Source material comes from the Library of Alexandria, sessions run through Mori under WALDO policy, and mastery summaries land in the audit trail, so training evidence documents understanding, not clicks.

Next step

Turn your next policy rollout into proven understanding.

A briefing watches MENTOR turn a source document into a mastery-gated learning session, live.