Security & Compliance

Built for regulated industries.

HIPAA. PCI-DSS. TCPA. HMAC webhooks. Audit trails. Compliance isn't a feature — it's the architecture.

HIPAA BAA

Available

PCI-DSS

Scoped

TCPA

Automated

SOC 2

In progress

AES-256

Audio storage

Compliance Coverage

Regulated industries. Enforced at the infrastructure layer.

HIPAA BAA

Platform & Enterprise
  • Business Associate Agreement available on Platform and Enterprise plans
  • All call audio stored in Backblaze B2 with AES-256 encryption
  • PHI redacted from voicemail extractions before storage
  • Access logs retained 7 years

PCI-DSS Scope Reduction

All Plans
  • Card numbers are never spoken into call transcripts
  • Extraction layer masks PANs before storage
  • Reduces your PCI scope for phone-based payment flows
  • No cardholder data environment (CDE) on WFW infrastructure

TCPA Automation

All Plans
  • Automated calling restricted to 8am–9pm local time for the recipient
  • TCPA compliance rules enforced at the Tool Gateway layer
  • Bots cannot override calling-hours restrictions
  • Do-not-call list scrubbing before any outbound campaign

HMAC-Signed Webhooks

All Plans
  • Every outbound webhook signed with HMAC-SHA256
  • Timestamp included in every signature to prevent replay attacks
  • 5-minute timestamp window — older requests rejected automatically
  • Your systems verify signatures before processing any payload

SOC 2 Roadmap

In Progress
  • SOC 2 Type II audit in progress — expected Q3 2025
  • Role-based access controls implemented and auditable
  • Audit logging enabled for all privileged operations
  • Encryption at rest and in transit across all services

Fair Housing Compliance

Real Estate Vertical
  • Prohibited phrases filtered at the Tool Gateway
  • Agents cannot make statements that violate Fair Housing Act guidelines
  • Steering language detection built into the extraction layer
  • Audit log of filtered phrases available for compliance review

Legal Vertical

For Legal Teams

ABA Model Rules of Professional Conduct are enforced at the infrastructure layer — not left to individual agent prompts.

What WFW Enforces Automatically

  • No unauthorized practice of law — agents cannot provide specific legal advice
  • Mandatory attorney disclosure on first interaction
  • Conflict screening before any case discussion begins
  • Privilege warning before intake questions are asked

What You Configure

  • Jurisdiction-specific rules and disclosures
  • Practice area intake form questions
  • Referral workflows for out-of-scope cases
  • Fee structure and consultation scheduling

Data Architecture

Where your data lives — and who can see it.

Every data store is scoped to your client ID. Row-level multi-tenancy is enforced in every query — there is no code path that returns another tenant's data.

StoreContains

PostgreSQL (Neon)

Relational DB

Agent configs, call metadata, extraction results

Backblaze B2

Object Storage

Call audio files

Upstash Vector

Vector DB

Knowledge base embeddings

Redis (Upstash)

Event Bus / Cache

Event bus, idempotency keys, rate limits

Row-level multi-tenancy

Every database row is scoped to clientId. No cross-tenant data leakage is possible — isolation is enforced at the query layer, not the application layer.

Auth Architecture

Zero-cookie, zero-CSRF bot surface.

The human dashboard and the bot API use entirely separate authentication models. There is no shared session state between the two surfaces.

Human Surface

  • JWT session cookies (NextAuth) — secure, httpOnly
  • 24-hour expiry with automatic rotation
  • CSRF protection via SameSite=Strict cookie policy
  • OAuth 2.0 social login (Google, GitHub) available
  • MFA support via TOTP authenticator apps

Bot Surface

  • Bearer tokens only — no cookies, no CSRF surface
  • Scoped API keys — agents:read, agents:write, calls:read
  • OAuth 2.1 client credentials flow
  • Token binding to clientId — no cross-tenant token reuse
  • API keys stored as bcrypt hashes — WFW staff cannot read your keys

Audit Trail

Every action. Every actor. Traced.

Every API call is logged with actor identity, timestamp, IP, user agent, and response status. Every automated action is traceable to the human-approved policy that authorized it.

“Why did my bot do that?” — one click in the dashboard shows you the policy that triggered the action, the call that caused it, and the exact response returned.

  • API audit log — every bot call with full context
  • Policy execution log — every automated action traced to its policy
  • Human action log — every dashboard interaction
  • Log retention: 90 days default, 7 years with HIPAA BAA

Audit Log — Recent Events

sa_bot_acme

POST /v2/agents

201 Created

2m ago

Policy: recall_reminder

wfw_send_confirmation_sms

Executed

4m ago

user@practice.com

Updated system prompt v4

Approved

12m ago

sa_bot_acme

GET /v2/calls?agent_id=...

200 OK

18m ago

Request the Security Package

Get the completed security questionnaire, DPA, HIPAA BAA, and subprocessor list in one email.

Delivered within 1 business day. All documents signed and ready to execute.

Questions about our compliance posture? Email security@workforcewave.com