Data Privacy

Twig AI's commitment to protecting your data and ensuring privacy compliance.

Privacy Principles

  1. Data Ownership: You own your data completely

  2. No Training: Your data never trains our models

  3. Encryption: Data encrypted at rest and in transit

  4. Isolation: Complete organization data isolation

  5. Transparency: Clear data usage policies

Data Collection

What We Collect

User Data:

  • Email address (required)

  • Name (optional)

  • Organization affiliation

  • Usage patterns (anonymized)

Content Data:

  • Documents you upload

  • Data source connections

  • Knowledge base articles

  • Conversation history

  • Agent configurations

Technical Data:

  • API request logs

  • Performance metrics

  • Error logs

  • System diagnostics

What We DON'T Collect

❌ Payment card details (handled by Stripe) ❌ Passwords (only hashed versions) ❌ Unnecessary personal information ❌ Data from other organizations

Data Storage

Storage Locations

Primary Region: US-East (AWS) Available Regions (Enterprise):

  • US-West

  • EU (Frankfurt)

  • Asia-Pacific (Singapore)

  • Custom regions on request

Encryption

At Rest:

  • AES-256 encryption

  • Encrypted database fields

  • Encrypted file storage (S3)

  • Key management via AWS KMS

In Transit:

  • TLS 1.3

  • HTTPS only

  • Perfect forward secrecy

  • Strong cipher suites

Data Access

Who Can Access Your Data?

Within Your Organization:

  • Users with appropriate permissions

  • Based on role and group membership

  • Audit-logged access

Twig AI:

  • Support team (with your permission only)

  • For troubleshooting specific issues

  • Fully audit-logged

  • Never for training or other purposes

Third Parties:

  • ❌ Never shared

  • ❌ Never sold

  • ❌ Never used for advertising

  • ⚠️ LLM providers process queries (see below)

LLM Provider Processing

When you use Twig AI:

  • Queries sent to LLM providers (OpenAI, etc.)

  • Zero-retention policy enforced

  • Data not used for training

  • Enterprise agreements in place

Control:

Privacy Controls

Private Data Mode

Restrict agents to organization-only data:

Effect:

  • No external knowledge used

  • Only your uploaded/connected data

  • Highest privacy

  • More controlled responses

Data Classification

Tag data by sensitivity:

PII Detection

Automatically detect and handle PII:

Example:

Data Retention

Default Retention

Data Type
Retention Period

Interactions

90 days

Analytics

2 years (aggregated)

Audit Logs

1 year

Documents

Until deleted

User Accounts

Until deleted

Custom Retention (Enterprise)

User Rights

Right to Access

Users can request all data we hold:

  1. Settings → Privacy → Request Data Export

  2. Receive email within 30 days

  3. Download complete data archive

Includes:

  • Profile information

  • Interaction history

  • Analytics data

  • Audit logs

Right to Deletion

Users can request data deletion:

  1. Settings → Privacy → Delete My Data

  2. Confirm deletion

  3. Processed within 30 days

Effect:

  • User account deleted

  • Personal data removed

  • Interactions anonymized

  • Analytics aggregated (no personal link)

Right to Portability

Export your data in standard formats:

  • JSON (machine-readable)

  • CSV (spreadsheet)

  • PDF (human-readable)

Compliance

GDPR (EU)

✅ Data processing agreements ✅ Lawful basis for processing ✅ Right to access, deletion, portability ✅ Data protection officer designated ✅ Privacy by design ✅ Breach notification (72 hours)

CCPA (California)

✅ Right to know what data is collected ✅ Right to delete ✅ Right to opt-out of sale (we don't sell data) ✅ Non-discrimination

HIPAA (Healthcare)

Available for Enterprise: ✅ Business Associate Agreement (BAA) ✅ Encrypted data storage ✅ Access controls ✅ Audit logging ✅ Breach notification

SOC 2 Type II

✅ Annual audit ✅ Security controls ✅ Availability controls ✅ Confidentiality controls ✅ Report available on request

Privacy-First Features

Data Minimization

We only collect what's necessary:

Purpose Limitation

Data only used for stated purposes:

  • Provide AI services

  • Improve platform

  • Support customers

  • Comply with legal obligations

NOT used for:

  • Marketing to non-customers

  • Selling to third parties

  • Training public models

Anonymization

Analytics use anonymized data:

Incident Response

Data Breach Protocol

If breach occurs:

  1. Detection: Identify and contain (< 1 hour)

  2. Assessment: Evaluate scope and impact (< 6 hours)

  3. Notification: Inform affected users (< 72 hours)

  4. Remediation: Fix vulnerability

  5. Review: Post-mortem and improvements

User Notification

You'll be notified if:

  • Data breach affects your account

  • Unauthorized access detected

  • Security settings changed

  • Suspicious activity observed

Privacy Settings

Organization-Level

User-Level

Users can control:

  • Data sharing preferences

  • Analytics inclusion

  • Email communications

  • Third-party integrations

Best Practices

1. Classify Your Data

✅ Identify sensitive data ✅ Apply appropriate controls ✅ Restrict access ✅ Audit regularly

2. Enable Private Data Mode

✅ For sensitive use cases ✅ For compliance requirements ✅ For controlled environments

3. Regular Audits

✅ Review access logs quarterly ✅ Verify permissions annually ✅ Remove inactive accounts ✅ Update security policies

4. Train Your Team

✅ Privacy awareness training ✅ Data handling procedures ✅ Incident reporting ✅ Secure practices

Next Steps

Last updated