Enterprise Privacy Architecture

Anonymity that's
mathematically
enforced.

PulseFlow doesn't ask employees to trust a policy. We bake anonymity into the architecture itself — through Rule-of-5 Redaction and Response Decoupling — so even we can't trace who said what.

k-Anonymity Guaranteed
Zero Identity Tokens Stored
TLS 1.3 + AES-256
GDPR Privacy-by-Design
Zero Data Commercialization
N ≥ 5
Redaction threshold
0
Identity–payload links
256-bit
Encryption at rest
100%
Privacy-by-design
Security Foundations

Four layers of privacy protection

Privacy guarantees that survive even insider access — because the data architecture makes re-identification mathematically impossible.

Rule-of-5 Redaction
Any segment with fewer than 5 respondents is automatically suppressed. No admin can bypass or lower this hardcoded threshold.
See live demo →
Response Decoupling
Identity tokens are verified and purged at the API proxy in volatile memory — they never reach the feedback database.
View architecture →
Bank-Grade Encryption
TLS 1.3 with HSTS enforces all transit. AES-256 with tenant-isolated keys secures all data at rest, with automated rotation.
Read specs →
Zero-Sell Ethics
We never sell, rent, or commercialize client data. Feedback is never passed to public AI models for training — not now, not ever.
Read FAQ →
k-Anonymity Protection

Rule-of-5
Redaction Engine

In small teams, responses can be reverse-engineered by elimination. PulseFlow's aggregation layer enforces a hardcoded minimum of N ≥ 5 respondents before any metrics are surfaced.

Automated suppression — scores, sentiment, and verbatim comments are all withheld for any cohort under the threshold.

Zero admin override — HR Super-Admins and executives have no mechanism to lower or bypass the N=5 guardrail.

Cross-filter defense — combining multiple filters (Department + Tenure + Location) recalculates cohort size in real-time before any data is shown.

aggregation-engine.pulseflow
Live
Select cohort to simulate:
Product Department — Engagement Report
84 / 100
Connect & Trust88%
Autonomy & Growth80%
Wellbeing & Recognition83%
✓ Aggregated from 12 responses. Individual identification statistically impossible.
[ Redacted — Cohort N < 5 ]
Analytics are withheld to protect individual identities within this segment.
RULE-OF-5 ENFORCED
Received: 3 responses
Required: ≥ 5 to display
Zero-Trace Ingestion

Response Decoupling

Standard surveys store user emails next to responses in the same database row. PulseFlow separates identity from payload at the API boundary — before anything is written to disk.

Survey Submission
Employee opens a unique magic link and submits their responses. Request carries both identity token and answer payload.
Token: #a8c2e1f...
Payload: {score: 9, comment:...}
Proxy Validation
Edge proxy validates token authenticity to prevent duplicate submissions — without persisting the identity to any log storage.
Auth: PASS
Token status: CONSUMED
Written to log: NONE
Identity Purge
Token, IP address, and all request metadata are stripped from memory. The payload is re-encrypted with a tenant analytics key.
IP header: STRIPPED
Auth token: PURGED
Metadata: ZEROED
Anonymous Storage
Only the unlinked, encrypted feedback vector reaches the analytics database. Re-identification is architecturally impossible.
DB entry: isolated vector
Traceable keys: ZERO
Foreign keys: NONE
Standard platforms
Store user_id and response in the same row, then set is_anonymous = true. That flag can be queried away by any DB admin with direct access.
PulseFlow architecture
Identity and payload are processed in separate memory spaces. By the time feedback is persisted, there is no foreign key to join, no metadata to cross-reference, and no token to trace.
Technical Controls

Security specifications

A full breakdown of PulseFlow's technical controls, encryption standards, and data handling protocols for enterprise evaluation.

Security Domain Technical Standard Protection Assurance
Data in Transit TLS 1.3 / HSTS · 4096-bit RSA All traffic between clients, proxies, and databases is encrypted with TLS 1.3 and Perfect Forward Secrecy. HTTP connections are rejected.
Data at Rest AES-256 · Tenant key isolation Storage disks, backups, and log streams are AES-256 encrypted. Each client's data uses isolated tenant keys with automated rotation.
Respondent Anonymity Rule-of-5 · Decoupled proxies Hardcoded N=5 threshold. Identity tokens are purged in volatile proxy memory before any payload reaches persistent storage.
GDPR Alignment Article 25 Privacy-by-Design Data minimization and purpose limitation enforced architecturally. Auth records kept minimal and transient per right-to-erasure obligations.
Data Commercialization Zero monetization policy Client feedback is never sold, rented, or licensed. We do not train public AI LLMs on your employees' responses — under any circumstances.
Security Contact sidharth.technologies@gmail.com Direct channel for DPA requests, vendor security assessments, and vulnerability disclosure. Responses within 24 hours.
FAQ

Security & compliance questions

Questions we hear from HR leaders, IT security teams, and enterprise procurement — answered plainly.

Have a DPA request?
Email our security team
sidharth.technologies@gmail.com
No. Neither HR admins, company leadership, nor PulseFlow engineers can link a specific submission back to an individual. Identity tokens are purged at the proxy layer before data persists, and the Rule-of-5 engine redacts analytics for any segment too small to hide in.
Their responses are still collected and included in higher-level company-wide aggregations (where N≥5), so no feedback is lost. However, segment-level scores and comments for that specific group are suppressed to protect individuals.
Most survey tools store email and response in the same row with is_anonymous = true. Any DBA can bypass that flag. PulseFlow's decoupling strips identity headers in volatile memory before payload persistence — there is no record to undelete.
Never. Client feedback, verbatim comments, and engagement data remain strictly within your workspace. We do not pass any survey content to public AI language models for training — not now, not ever.
Email sidharth.technologies@gmail.com with your DPA requirements or vendor assessment questionnaire. We aim to respond within 24 hours.
Our architecture is designed around GDPR Article 25 (Privacy-by-Design). Data minimization is enforced structurally — we collect only what is needed, keep auth records minimal and transient, and support right-to-erasure obligations for employee authentication data.