CEP Validation Study
The Study
We're validating a framework for measuring personality through conversation — domain by domain, not as a single global trait. The insight that drives everything: how you think about one thing is different from how you think about another. We're measuring that variance and using it to configure agents that complement how you actually think.
Closed Beta
The Hypothesis
Personality-aware, intent-conditioned agents produce better conversation outcomes than generic agents — when personality is measured through conversation and agents are configured to complement the user's cognitive style.
The Measurement Framework
Cognitive Expression Profile
Our method is behavioral observation: what you say during structured conversations, not what you claim about yourself.
We measure multiple cognitive dimensions — how someone engages with new ideas, how they approach evidence, how they handle disagreement and uncertainty. Each dimension is observable in how a person actually engages, not how they'd describe themselves.
The framework draws on decades of validated personality research. Our contribution is the measurement method: deriving these dimensions from conversation rather than self-report, and measuring them domain-specifically rather than as global traits.
Why Domain-Specific Matters
Personality isn't universal. How you think about climate policy differs from how you think about finance.
CEP measures domain-specific cognitive patterns — and that variance is itself the finding. Nobody's published domain-specific personality measurement at scale from conversational observation.
What We're Testing
- – Can we reliably measure cognitive dimensions through conversational interview?
- – Do behavioral scores from debates converge with interview-derived profiles?
- – Does personality-aware agent matching improve goal achievement vs. a baseline?
- – Does cognitive style vary meaningfully across domains?
How You Participate
Conversational Interview
5–8 minutes
A short conversational interview about how you engage with a topic you choose. No right answers — we're observing how you think, not testing what you know.
Structured Debates
Engage with AI personas on topics you care about. Share your perspective, they push back, you respond. The system builds a picture of how you think from how you engage.
Profile Complete
Ongoing
Agents are configured to be complementary to your cognitive style. The more domains you engage with, the richer the cross-domain pattern becomes.
What Comes Out of It
You don't see a score. You see what's interesting about how you think.
- – Domain-specific cognitive insight
- – Cross-domain patterns if you complete multiple domains
- – Personalized agents calibrated to your style
- – Contribution to research on personality-aware AI
Research Studies
01
CEP Measurement Validity
Do interview-derived cognitive profiles converge with behavioral scores from debates and self-report? Testing the core measurement validity of the CEP framework.
02
Intent Routing Effectiveness
A controlled study of personality-aware agent matching. Does calibrating to cognitive style improve conversation outcomes?
03
Cross-Domain Patterns
How does cognitive style vary across domains? Measuring whether domain-specific profiles diverge meaningfully from global trait scores.
Who This Is For
- – People who want real engagement with ideas
- – People curious about how they think
- – People willing to engage, not just consume
Interested in Collaboration?
Researchers, educators, platform builders. Let's work together.
Get in Touch