How it works

The science under the warm coach voice.

We will not insult your intelligence. Here is what the assessment measures, how it scores, and where we are confident versus where we are still learning.

What we measure

The Discover assessment combines five validated instruments and one custom layer.

  1. Ikigai four circles. Twenty items across love, skill, world need, paid for. Custom, validated against ~5,000 completions in beta.
  2. Big Five personality. NEO PI inspired short form covering openness, conscientiousness, extraversion, agreeableness, neuroticism.
  3. Schwartz values. Ten value types including autonomy, mastery, impact, security, legacy.
  4. Working preferences. Solo vs team, structure vs ambiguity, depth vs breadth, risk tolerance, time horizon.
  5. Constraints. Time per week, capital, location flexibility, family obligations, runway.
  6. Skills inventory. Self report grid plus optional LinkedIn import.

How we score

Scoring is deterministic and reproducible. Every assessment row stores the rubric version that scored it. We use classical test theory for personality and values, weighted composites for ikigai dimensions, and an interpretable archetype matcher.

The archetype matcher computes a weighted similarity between your scores and 12 archetype fingerprints. We show you which dimensions moved your score, in plain language. No black boxes.

Where we are confident

We are confident in the Big Five and Schwartz values portions, because these instruments have decades of literature behind them; in the deterministic scoring, because it is pure math; and in the archetype library, which reflects real founder data we have cross checked.

Where we are still learning

We are less certain about the precise weighting between ikigai dimensions and Big Five for each archetype. Those weights move slightly each rubric version as we collect more outcome data.

How we use AI

The Profile narrative is generated by an LLM (Google Gemini 3 Pro by default, Anthropic Claude as a fallback) with retrieval grounding on your actual scores. We do not invent data. The LLM never sees other users' data. Every flagship prompt has an offline eval suite of 50 to 200 graded examples that runs on every release.

What we don't do

We do not give legal, medical, or financial advice. We do not claim certainty about your future. We do not flatter. We do not sell your data, full stop.

Reassessment

We prompt you to retake a short form of the assessment every 30 to 90 days. Your Profile evolves. We show drift over time as a chart on your dashboard. The longitudinal Profile is the asset.