Your Data, Your Privacy

Privacy matters — that's why PAICE (People + AI Collaboration Effectiveness) follows Privacy By Design principles, and has from the very beginning. When you share your work and collaboration patterns with an assessment tool, you deserve to know exactly what happens to that information.

We continue to be completely transparent about what data PAICE collects, how it's processed, what's stored, and what's not.

The Short Version

Your conversation is not permanently stored. It's processed in real-time to generate your score, then discarded. We retain only:

• Anonymized scores (for our validation research)
• Aggregate, anonymized patterns (for experience improvement)
• Any feedback you explicitly and voluntarily provide

Your results belong to you. We don't share them with employers, other users, or third parties without your explicit consent.

We don't sell user data. Period.

You can choose hardware-level privacy. Our Confidential Mode routes your assessment through Trusted Execution Environments (TEEs) where not even we can access your conversation. This feature is now available to paid plans only, and is detailed in our newest whitepaper Confidential Mode: Trusted Execution Environments for AI Collaboration Assessments. But this post is about the basics available to everyone via our free assessment in Standard Mode.

Now let's dive into the details.

What Happens During Your Assessment

Multi-Model, Same Privacy

PAICE uses AI models from multiple providers — Anthropic (Claude), Google (Gemini), and OpenAI (ChatGPT) — to power different functions within each assessment. Different models handle conversation, evaluation, and detection tasks based on what each does best, with automatic failover across providers for reliability.

Regardless of which models are active in your session, the privacy protections are identical:

1. Your messages are processed via AI provider APIs — routed through whichever models are handling your session
2. A proprietary middleware layer strips potential PII before any language model sees your content
3. PAICE observes interaction patterns in real-time
4. Behavioral patterns are analyzed to generate your score
5. The conversation is discarded after assessment completion

Important: Your conversation content is not stored in PAICE's database. We process it in real-time, extract behavioral patterns, and then it's gone.

What We Extract

From your conversation, we extract:

• Behavioral patterns — how you frame requests, respond to outputs, iterate, verify, and recover from errors
• Collaboration metrics — effectiveness indicators across 20 parameters that roll up to our five core dimensions (Performance, Accountability, Integrity, Collaboration, Evolution) and produce your PAICE Score™
• Statistical features — conversation length, iteration count, verification frequency

We do not extract or store:

• The specific content of your work
• Proprietary information from your task
• Personal details mentioned in conversation
• Identifiable information about your employer or role

What We Store

Your Score and Results

We store your assessment results with anonymized identifiers — one-way cryptographic hashes that cannot be reversed:

• Your overall PAICE Score™ (0–1000 scale)
• Your dimensional breakdown across all five dimensions
• Your tier classification
• Your confidence interval
• Timestamp of assessment

This data is stored to:

• Allow you to download your results as a PDF
• Enable comparison if you retake the assessment
• Power your History page for longitudinal progress tracking

You can request deletion of this data at any time by contacting us.

Aggregate, Anonymized Patterns

For validation research, we retain:

• Statistical distributions of scores
• Patterns in dimensional profiles
• Aggregate behavioral trends
• Population-level insights

This data is:

• Completely anonymized (no way to link it back to you)
• Aggregated across many users
• Used only for research and validation
• Never shared in a way that could identify individuals

Feedback You Provide

If you choose to provide feedback after your assessment, we store:

• Star ratings of accuracy
• Text feedback about your experience
• Suggestions for improvement

This feedback is:

• Stored with your session ID (not your personal identity)
• Used to improve the assessment
• Reviewed by our team
• Never shared publicly with attribution

Email Addresses (Optional)

If you provide an email address:

• Your email is encrypted at rest using AES-128 encryption
• It's used only to send you direct follow-up, support, and major project updates
• You can unsubscribe from future communications
• You can request deletion at any time (within 30 days for GDPR or 45 days for CCPA)
• Most importantly, your email is not linked to your score or session experience in any way

We do not:

• Sell email addresses
• Share them with third parties
• Use them for marketing without consent
• Send unsolicited communications

What We Don't Store

Your Conversation Content

We do not permanently store:

• The messages you send during assessment
• The AI's responses to you
• The specific content of your work task
• Any proprietary or sensitive information you discuss

Why not? Because we don't need it. We extract behavioral patterns in real-time, and storing conversation content would create unnecessary privacy risk.

Personal Identifying Information

We do not collect or store:

• Your name (unless you provide it in feedback)
• Your employer or organization
• Your location or IP address (beyond basic website analytics)
• Your demographic information
• Your professional details

PAICE is designed for personal development, not surveillance. We don't need to know who you are to help you improve. You're safe to learn at your own speed, and to try new things without wondering who might be watching.

Third-Party Data Sharing

AI Model Providers

Your conversation is processed through APIs from the AI providers active in your session. This means:

• AI providers receive your conversation as part of API processing
• PAICE's PII middleware strips identifiable information before content reaches any provider
• Providers may retain data according to their respective commercial terms
• No provider trains on your assessment data — all are accessed via commercial API terms that prohibit training on customer inputs

Relevant provider terms:

• Anthropic Commercial Terms
• Google Cloud AI Terms
• OpenAI API Terms

No Other Third Parties

We do not share your data with:

• Employers or organizations
• Other assessment users
• Marketing platforms
• Data brokers
• Analytics companies (beyond basic, anonymized web analytics)

The only exception: If you explicitly choose to share your results (via PDF download, screenshots, social sharing, etc.), that's your choice and within your control.

Your Rights and Control

Access Your Data

You can:

• Download your results as a PDF immediately after assessment
• View your complete assessment history on your History page
• Request access to any stored data by contacting us

Delete Your Data

You can request deletion of:

• Your assessment results
• Your email address
• Your feedback

Contact us to request deletion. We process requests within 30 days (GDPR) or 45 days (CCPA).

Note: Aggregate, anonymized research data cannot be deleted because it's not linked to your identity.

Track Your Progress

Your History page provides longitudinal tracking of your development:

• Compare to your baseline: See how each dimension has changed over time
• Visualize trends: Track progress across multiple assessments
• Identify patterns: Understand which dimensions improve fastest and which need more work
• Derive insights: Get personalized recommendations based on your assessment history

Control Sharing

You decide whether to share your results:

• Download PDF for personal records
• Share on social media (optional)
• Discuss with colleagues or mentors
• Keep completely private

While we encourage you to be loud and proud about your PAICE Score™, we never share your results without your explicit action.

Security Practices

Data Protection

We implement defense-in-depth security:

• Encrypted data transmission (HTTPS/TLS 1.3)
• Encrypted data storage
• One-way cryptographic hashing for all user identifiers
• Access controls and authentication
• Regular security reviews
• Agentic browser detection to prevent automated assessment-taking (without fingerprinting or tracking you)

Limited Data Retention

We practice data minimization:

• Store only what's necessary
• Discard conversation content immediately
• Retain aggregate data only
• Provide deletion on request

No Surveillance

PAICE is designed for self-assessment, not monitoring:

• Individual users control their data
• Results are not shared without consent
• Assessment is voluntary
• Focus is on learning, not judgment

Organizational Use

When organizations use PAICE for team or cohort assessment:

• Individual scores are structurally protected — enterprise buyers receive cohort-level aggregations only (distributions, percentiles, trend lines), with no individual mapping
• This protection is architectural, not just policy — the system is designed to make individual identification from cohort data structurally impossible
• Individual consent is required for any data sharing
• Privacy protections are maintained or enhanced at every level

We will not:

• Retroactively share individual assessment data
• Change privacy practices without notice
• Use data in ways not disclosed upfront

Common Questions

"What if my employer wants my results?"

That's your decision, not ours. Your results belong to you. We will never share them with your employer without your explicit consent. Even when organizations use PAICE for cohort assessment, they receive only aggregate data — your individual score is not accessible to them.

"Can I take the assessment anonymously?"

Yes. It is anonymous by default. You don't need to provide any identifying information to take the assessment. The only optional identifier is an email address if you choose to provide one, but even then it is not linked to your score or session.

"What about AI provider data retention?"

Each AI provider has their own data retention policies for API usage. PAICE's PII middleware strips identifiable information before content reaches any provider, and all providers are accessed via commercial terms that prohibit training on customer data. For the strongest guarantee, consider our Confidential Mode which routes processing through TEE hardware where no external party can access your data during processing.

"What happens if PAICE is acquired or shuts down?"

We commit to:

• Providing notice before any ownership change
• Giving users the opportunity to delete their data
• Maintaining privacy commitments through transitions
• Deleting data if we shut down (except aggregate research data)

"I'm worried about proprietary information."

Don't include it in your assessment task. Choose a work challenge that doesn't involve trade secrets, confidential client information, or sensitive business data. Your conversation isn't stored, but it is processed through AI provider APIs. Again, if this concern is paramount, consider Confidential Mode for hardware-level protection.

The Bottom Line

Your privacy is not negotiable. We've designed PAICE with privacy as a core architectural principle, not an afterthought.

Your conversation is not stored. We process it in real-time, extract behavioral patterns, and discard the content.

Your results belong to you. We don't share them without your consent, and we don't sell your data.

You have control. You can access, export, track, or delete your data at any time.

You have options. Standard mode provides strong privacy by design and architecture. Confidential Mode provides hardware-enforced privacy that not even we can circumvent.

Privacy matters because trust matters. We're building PAICE to be worthy of your trust.

---

Have questions about privacy or data practices? Contact us

Read our full Privacy Policy and Terms of Service for complete details.

Recommended Reading

📖 Privacy & Security:

• Introducing Confidential Mode - Hardware-level TEE privacy option
• Privacy by Design: How PAICE Achieves Privacy Compliance - Technical architecture deep dive

📖 How PAICE Works:

• Will PAICE Support Other AI Models? - Multi-model architecture explained
• The Ethics of AI Collaboration - Broader ethical considerations