Data Flow Architecture
The path of an inbound call from connection through transcription, AI processing, secure storage, and configurable CRM sync.
- Incoming call, Telephony provider receives the call.
- Speech-to-text, Audio transcribed in real time.
- LLM processing, AI generates the response.
- Secure storage, Encrypted at rest (AES-256).
- Integrations, CRM sync (configurable).
Security Controls
- Encryption in transit, All data encrypted via TLS 1.2+ in transit between all services.
- Encryption at rest, AES-256 encryption for all stored data including recordings, transcripts, and metadata.
- Access controls, Role-based access control (RBAC) with SSO/SCIM support on enterprise plans.
- Network security, Cloud infrastructure with WAF, DDoS protection, and network segmentation.
- Monitoring, 24/7 infrastructure monitoring with automated alerting.
- Incident response, Documented incident response plan with defined notification windows.
Subprocessor List
| Category | Provider | Purpose | Data Processed |
|---|---|---|---|
| Telephony | Twilio | Call handling and routing | Call audio, caller ID, metadata |
| Speech-to-Text | Deepgram | Real-time transcription | Call audio |
| LLM | OpenAI / Anthropic | AI response generation | Transcribed text, context |
| Cloud Hosting | AWS | Application and data hosting | All application data |
| Analytics | Internal | Call analytics and reporting | Aggregated call metadata |
AI Data Use Policy
Customer call recordings and transcripts are not used to train shared AI models. My AI Front Desk processes call data solely to deliver the AI receptionist service. Call data is not shared with LLM providers for model training. Enterprise customers can request details on data isolation and processing boundaries.
Frequently Asked Questions
Is call data used to train AI models?
No. Customer call recordings and transcripts are not used to train shared AI models.
Where is data stored?
Data is stored in AWS with AES-256 encryption at rest.
Who are your subprocessors?
Telephony: Twilio. STT: Deepgram. LLM: OpenAI/Anthropic. Hosting: AWS. Full list available in the procurement pack.
What security documentation do you provide?
We provide a security architecture overview, DPA, subprocessor list, retention policy, and AI data use policy. Available on request.
What happens if there is a security incident?
We follow a documented incident response plan with defined breach notification windows.