How Your Score Works
Your engagement score measures how effectively you interact with AI. It is an objective metric — the same formula applies to everyone — so you can track your growth over time and compare meaningfully with others.
The Score (0-100)
Your score is a weighted combination of six components, each grounded in established research on learning, cognition, and expert performance. The score is the same formula for every user — what changes is the feedback interpretation, which is personalized to your profile.
The Formula
Score = (Alignment × 0.30) + (Transitions × 0.20) + (Breadth × 0.15) + (Agency × 0.15) + (Sequence × 0.10) + (Verification × 0.10)
Each component is independently scored 0-100, then combined using these weights. The result is your overall score.
Task Alignment
30%Different tasks call for different engagement patterns. Writing a paper requires creative exploration and critical challenge. Studying for an exam requires tutoring and verification. This component measures how well your mode distribution matches the recommended pattern for YOUR specific task type.
Based on Person-Environment Fit theory (Edwards, 1986) — outcomes improve when behavior matches task demands.
Mode Breadth
15%How many of the 8 engagement modes did you use? Most tasks benefit from 4-6 different modes. Using only 2-3 modes suggests you're stuck in a pattern, even if those modes are used well. Range matters.
Based on Deliberate Practice (Ericsson et al., 1993) — experts develop by practicing the full range of skills, not just comfortable ones.
Agency Level
15%What proportion of your messages were in the Agency tier (Modes 5-8: Verification, Creative Expander, Critical Challenger, Problem Setter)? Higher agency means you were driving the thinking rather than receiving from AI. This is adjusted for what the task expects — some tasks legitimately need more Partnership or even Passivity modes.
Based on Bloom's Taxonomy (1956) and Self-Regulated Learning (Zimmerman, 2002) — higher-order cognitive engagement produces deeper learning outcomes.
Mode Dynamism (Transitions)
20%How dynamically did you move between engagement modes during the conversation? This looks at the variety of mode-to-mode transitions and how often the conversation shifts. Fluidly moving between modes — say, from asking a question to challenging the answer to verifying it — reflects adaptive engagement rather than staying in one gear the whole time. It is measured on this conversation alone, so it means something on your very first submission.
Based on Adaptive Expertise (Hatano & Inagaki, 1986) — flexibly switching approaches, rather than applying one routine, is a marker of expert rather than rote engagement.
Sequence Quality
10%Did you follow the recommended mode sequence for this task? Expert problem-solvers don't just use the right tools — they use them in the right order. For example, the writing path recommends: frame the assignment (Mode 8) before brainstorming (Mode 6) before drafting (Mode 2). Starting with drafting and working backward scores lower on this component.
Based on Problem-Solving Protocols (Schoenfeld, 1985) and Cognitive Apprenticeship (Collins, Brown & Newman, 1989) — expert sequences transfer across domains when learned explicitly.
Verification Habit
10%Did you verify AI's output at least once? Regardless of task type, checking AI's work is a foundational skill. This component rewards any amount of verification (Mode 5) rather than requiring a specific percentage. The question is simply: did you check?
Based on Automation Bias research (Parasuraman & Manzey, 2010) — humans systematically over-trust automated systems. Active verification must be a learned habit, not an assumed behavior.
What the Score Does NOT Measure
Not the quality of your work output. The score measures how you engaged with AI, not whether your essay was good or your code compiled. You could produce excellent work with a low engagement score (by letting AI do everything) or mediocre work with a high score (by engaging deeply with material you don't yet understand).
Not a judgment of you as a person. It measures one conversation against one task type. Your next conversation could score completely differently. The score is a snapshot, not a label.
Not a measure of AI tool quality. The score reflects YOUR engagement patterns, not how good ChatGPT, Claude, or Gemini's responses were.
Where Personalization Lives
The score formula is identical for everyone. But the feedback you receive is personalized based on:
- Your personality profile — feedback is calibrated to your Big Five traits and Need for Cognition
- Your engagement history — recommendations reference your patterns across all submissions
- Your archetype — growth paths are specific to your archetype (Delegator, Partner, Verifier, etc.)
- Feedback optimization — the system learns which feedback style works best for users with profiles similar to yours
References
Bloom, B.S. (1956). Taxonomy of educational objectives: The classification of educational goals. Longmans, Green.
Collins, A., Brown, J.S., & Newman, S.E. (1989). Cognitive apprenticeship: Teaching the crafts of reading, writing, and mathematics. In L.B. Resnick (Ed.), Knowing, learning, and instruction (pp. 453-494). Erlbaum.
Dweck, C.S. (2006). Mindset: The new psychology of success. Random House.
Edwards, J.R. (1986). An examination of competing versions of the person-environment fit approach to stress. Academy of Management Journal, 39(2), 292-339.
Ericsson, K.A., Krampe, R.T., & Tesch-Romer, C. (1993). The role of deliberate practice in the acquisition of expert performance. Psychological Review, 100(3), 363-406.
Parasuraman, R. & Manzey, D.H. (2010). Complacency and bias in human use of automation. Human Factors, 52(3), 381-410.
Schoenfeld, A.H. (1985). Mathematical problem solving. Academic Press.
Vygotsky, L.S. (1978). Mind in society: The development of higher psychological processes. Harvard University Press.
Zimmerman, B.J. (2002). Becoming a self-regulated learner: An overview. Theory Into Practice, 41(2), 64-70.