How is my data safe with Authava?
Authava protects your data with encryption, enterprise hosting, separation between customers, and a design that keeps most of your data in your own systems. Your conversations and documents are not used to train public AI models.
Authava was built by an experienced enterprise engineering team based in the United States.
Infrastructure runs in Tier-1 datacenters.
Below are some safeguards built into the platform.
Infrastructure Controls
Enterprise hosting
Authava runs in Tier-1 datacenters.
Encryption
Data sent between visitors and Authava is encrypted with TLS. Our primary databases and file storage are protected by our cloud providers' at-rest encryption.
Access controls
Administrative access is restricted to authorized staff.
Data Protection
No model training on customer data
Authava does not use your conversations or documents to train public AI models, and our AI providers' business terms state that data sent through their APIs is not used to train their models.
Minimal data storage
Customer data integrations are typically handled through APIs. In many cases, the bot retrieves information from your system rather than storing it.
Payment security
Authava does not store credit card data. Payments are handled through Stripe.
Application Design Controls
Tenant isolation
Authava is designed as a true multi-tenant system. Each customer runs in their own tenant, created in seconds. Bots are designed to access only their own tenant's data.
Closed knowledge scope
Bots are designed to answer from approved sources such as documents, Q&A knowledge, and APIs. They don't search the open internet; they only use sites you approve.
API-first integrations
Customer data typically remains in your own systems. Bots call your APIs when needed instead of requiring you to upload internal databases.
Permission model
Bot workspaces support role-based permissions (Read / Update / Delete / Admin) so teams can limit access appropriately.
Bot abuse protection
Rate limiting and anti-spam protections minimize the abuse of bots.
Process Controls
Document preparation (Managed Service)
For managed bots, the Authava team reviews and prepares documents before they are used by the bot. This can include removing PII and structuring information safely.
Secure configuration for sensitive operations
Risk-prone operations are typically configured by the managed service team to avoid unsafe setups.
Integration Security
Authentication integration
Bots can integrate with existing authentication systems (for example JWT-based authentication). Your system remains the credential store.
Enterprise LLM providers
By default, Authava uses OpenAI's GPT models hosted on Microsoft Azure. The default can change over time and can be overridden.
White-label email security
Email notifications optionally send using OAuth-based connections rather than simple DNS forwarding.
Frequently Asked Questions
Is my data encrypted?
Data sent to and from Authava is encrypted in transit (TLS), and our primary databases and file storage are encrypted at rest by our cloud providers.
Does Authava use my data to train AI models?
No. Authava does not use your conversations or documents to train public AI models.
Do I have to upload my customer database?
No. Customer data integrations are typically handled through your APIs. In many cases, the bot retrieves information from your system when it's needed rather than storing it.
Can other Authava customers see my data?
Each customer runs in their own tenant, and bots are designed to access only their own tenant's data.
Does Authava store credit card numbers?
No. Payments are handled through Stripe.