Data Privacy
TLDR
- We do not sell your data.
- Your data remains private and isolated from other Authava customers.
- Authava was built from the ground up with tenant isolation as a core part of the platform.
- Each bot has its own knowledge base, documents, configuration, and conversation history.
- Your data is not used to train other customers’ bots or public AI models.
- Your data is not published to the internet.
Your data belongs to your business
The information you provide to Authava is used to operate your bot and deliver the services you configured.
This may include:
- Website content
- Uploaded documents
- Knowledge-base articles
- Bot instructions
- Conversation history
- Customer or lead information
- Data retrieved through integrations
- Reports and analytics
Authava does not sell your business data or repurpose it for another customer’s bot.
Your data is isolated by tenant
Authava was built from the ground up as a multi-tenant platform.
Each customer operates within a separate tenant. A unique tenant ID is applied throughout the platform and is foundational to how Authava stores, retrieves, and processes data.
This tenant separation applies to:
- Knowledge bases
- Documents and website content
- Bot instructions and configuration
- Conversation history
- Integrations
- Customer records
- Reports and analytics
A bot can access only the information assigned to its own tenant. One customer’s bot cannot retrieve information belonging to another customer.
Your data does not train other bots
Each Authava bot has its own knowledge, instructions, configuration, and conversation history.
Your documents and conversations are not placed into a shared training pool. They are not used to train another customer’s bot or improve answers for another business.
Information supplied to your bot is used only to provide the services configured for your tenant.
Your data is not published to the internet
Adding information to Authava does not publish it as a webpage or make it searchable through the open internet.
Authava bots answer from approved sources, including:
- Content supplied by your business
- Approved website pages
- Uploaded documents
- Configured question-and-answer content
- Authorized APIs and integrations
Information intentionally provided through a public-facing bot may, of course, be included in that bot’s answers. You control what information is made available to the bot and how the bot is configured to use it.
Enterprise AI providers
Authava uses enterprise AI model providers, such as Microsoft Azure OpenAI and other enterprise services.
These services are intended for business applications and provide data-handling protections. Customer prompts, responses, and documents sent through these enterprise services are not used to train public foundation models.
For advanced deployments, Authava can also integrate with a customer-hosted language model.
Customer data can remain in your systems
Authava is designed around API-based integrations.
In many cases, customer information does not need to be copied into Authava. Instead, the bot retrieves the information it needs from your existing system through an authorized API.
This helps your existing system remain the primary source of truth while allowing the bot to perform approved tasks.
You control what the bot can access
You decide:
- Which information is added to the bot
- Which websites or documents it may use
- Which systems it may connect to
- Which actions it may perform
- Who may access its workspace
- What information it may provide to users
Authava workspaces also support role-based permissions, allowing businesses to limit who can read, update, delete, or administer bot content.
When might two bots have overlapping information?
Tenant isolation does not mean two bots can never contain or produce similar information. For example, two finance bots asked “What is a stock?” may produce similar answers because the underlying concept is common knowledge.
Two bots may overlap when:
- Both bots are configured to read the same public webpage.
- Both bots are authorized to call the same API.
- The same document is intentionally uploaded to multiple bots, such as a public industry guide or product manual.
- Two businesses provide similar information independently.
- Two bots answer the same general-knowledge question.
This does not mean one bot accessed the other bot’s knowledge base, documents, conversations, or customer data. Each bot still retrieves information only from the sources authorized for its own tenant.
The bottom line
Your Authava bot is built for your business.
Your knowledge, documents, conversations, integrations, and customer information remain separated from other Authava customers. They are not used to train other bots, placed into a shared knowledge base, or published to the internet.
Your data stays with your tenant and is used only to operate the services you have configured.
See also: