Features
Secure Unmasking: Turn Protected Responses Into Usable Answers
Our Reveal Technology automatically restores original values in AI responses—for authorized users only, with complete access controls. The feature that transforms masked outputs from unusable placeholders into actionable intelligence.
The Usability Problem with Masking
Every AI privacy tool on the market masks sensitive data. They have to. But when the AI responds, those responses contain the same masked tokens. And that's where they all fail.
Unusable AI Responses
You asked: "What treatment should this patient receive?" You got: "[PATIENT_A] should receive [MEDICATION_1] at [DOSAGE_1] and follow up with [DOCTOR_A] on [DATE_1]."
This is useless. The clinician can't act on a response where every specific is hidden.
Manual Reconstruction
Workflow without Reveal: 1. Read masked response 2. Look up [PATIENT_A] in your notes 3. Look up [MEDICATION_1] in your notes 4. Look up [DOSAGE_1] in your notes 5. Look up [DOCTOR_A] in your notes 6. Look up [DATE_1] in your notes 7. Reconstruct the full sentence 8. Actually use the information
10-15 minutes of work per response. Security creates unsustainable friction.
Tool Abandonment
What users do: "This privacy tool makes AI useless. I'll just paste directly into ChatGPT." What happens: Complete security bypass. All protection wasted.
When security creates too much friction, people work around it.
Reveal Technology Solves This
Secured AI's Reveal Technology automatically restores original values in AI responses—but only for authorized users, with full access controls. Users get usable responses. Security gets control and visibility.
Benefits of Secure Unmasking
Instantly Usable AI Responses
No more manual reconstruction. Responses are automatically de-obscured with original values restored. Users can act on AI outputs immediately.
Zero reconstruction time for authorized users
Granular Access Control
Control reveal permissions by user role, data type, or context. Admins see everything; analysts might see names but not SSNs; some users see nothing.
Per-data-type, per-role permission configuration
Complete Activity Tracking
Every reveal operation is tracked: who, what, when, and why. Compliance teams can prove exactly who accessed what sensitive data.
Activity records with user identity, timestamp, and specifics
Session-Scoped Security
Token-to-value mappings are scoped to individual sessions. You can't use tokens from one session to reveal data from another.
Cryptographic session isolation
Configurable Retention
Choose how long reveal capability persists. Default: session only. Compliance option: extended retention for regulatory purposes.
Retention policies from immediate destruction to 7+ years
Seamless User Experience
For authorized users, reveal is automatic. They don't see tokens—they see the real response. Security is invisible.
Transparent reveal for authorized sessions
How Secure Unmasking Works
Response Reception
Input
LLM response containing masked tokens
Control
Capture response before presenting to user
Output
Response queued for reveal processing
Token Identification
Input
Masked response
Control
Scan for token patterns matching session's token registry
Output
List of tokens to potentially reveal
Permission Verification
Input
Token list + user identity
Control
Check each token against reveal policy: User's role; Data type of the token; Session context; Time-based rules
Output
Authorized tokens vs. denied tokens
Value Restoration
Input
Authorized tokens + encrypted mapping registry
Control
Decrypt mapping for authorized tokens only: Retrieve encrypted mapping; Decrypt with session key; Replace token with original value
Output
Partially or fully revealed response
Activity Recording
Input
Reveal results
Control
Create operation record: User identity; Timestamp; Session ID; Tokens revealed; Tokens denied
Output
Compliance-ready activity record
Response Delivery
Input
Revealed response + denied tokens (if any)
Control
Present to user: Authorized tokens show original values; Denied tokens show token placeholder
Output
User receives appropriately revealed response
Reveal Permission System
Reveal permissions determine who can see what. Configure at the organization, team, or individual level.
By Role
By Data Type
By Context
Activity Tracking for Compliance
Every reveal operation is tracked and recorded. These records support HIPAA access documentation, SOC 2 requirements, and internal security reviews.
User identity
Who performed the reveal operation
Timestamp
When the reveal occurred (UTC precision)
Session context
Which session the reveal occurred in
Data type
Category of the revealed data
Access policy
Permission level applied to the operation
Export
SIEM-compatible formats (JSON, CEF, Syslog)
Retention
Configurable from 30 days to 7+ years
Search
Query by user, date range, data type, session
Alerting
Real-time alerts on unusual reveal patterns
Technical Specifications
Reveal Performance
Security
Mapping encryption
AES-256-GCM
Key management
HSM-backed, per-session
Session isolation
Cryptographic binding
Data integrity
Cryptographic integrity verification
Compliance
HIPAA
Access documentation, minimum necessary
SOC 2
Access control, activity tracking
GDPR
Data access logging, right to access
CCPA
Access request documentation
Secure Unmasking in Action
Healthcare
Scenario:
Clinician reviews AI-generated patient summary
Reveal:
Patient names, diagnoses, medications, providers revealed
Tracking:
"Dr. Chen revealed [PATIENT_A] data at 2:34 PM for treatment planning"
Compliance:
HIPAA access documentation satisfied
Financial Services
Scenario:
Advisor reviews AI-prepared client meeting notes
Reveal:
Client names, account details, portfolio values revealed
Tracking:
"J. Park revealed [CLIENT_A] financial data at 9:15 AM"
Compliance:
Fiduciary duty documentation maintained
Legal
Scenario:
Attorney reviews AI case summary
Reveal:
Party names, case numbers, settlement amounts revealed (for authorized attorneys)
Tracking:
"Partner reviewed [CASE_123] details at 4:00 PM"
Compliance:
Privilege access documented
Frequently Asked Questions
If anyone can reveal, how is this secure?
What if someone screenshots revealed content?
Can I disable reveal entirely?
How long are reveal mappings retained?
Can I require approval for certain reveals?
What happens if a user's role changes?
Related Features
Real-time PII/PHI Detection
Detection identifies what needs to be protected before masking and reveal can work.
Context-Preserving Masking
Semantic tokenization creates the tokens that Reveal restores—while maintaining AI accuracy.
Reveal Technology Deep Dive
Learn more about the architecture and unique capabilities of our Reveal Technology.