Skip to main content

AI models and privacy

Overview​

This page explains how Business App AI Employees—including AI Chat Receptionist, AI Voice Receptionist, and custom AI Employees—produce answers and what we disclose about their underlying AI models.

Why is this important?​

  • You can make better updates to your content and capabilities when you know how answers are generated.
  • Clear expectations reduce confusion about model names, versions, or how they change over time.

What’s included​

  • Where AI Employees get information
  • Model disclosure policy
  • Guidance to improve responses
  • Frequently asked questions

How AI Employees generate answers​

AI Employees use a large language model (LLM) to draft responses based on:

  • Your configured capabilities. See Configure capabilities.
  • Your knowledge base content, including uploaded files and website sources. See Knowledge base.
  • Conversation context from the active chat, call, or task.
warning

Do not add sensitive or unnecessary personal data to your knowledge base. Only include information you want the assistant to use in replies.

Model disclosure policy​

  • We do not publish the specific LLM provider or model version used in production.
  • Model options may change over time to improve quality, safety, and reliability.
  • AI Chat Receptionist and AI Voice Receptionist automatically use a supported model. AI Employees let you choose from the supported models listed in Business App (the provider name is not shown).
  • The base LLM infrastructure is shared, but your account data is isolated. Your content is not visible to other businesses.

Improve response quality​

  1. Review your knowledge base for accuracy and completeness.
  2. Keep capability instructions short, specific, and action‑oriented.
  3. Test common customer questions and adjust wording where answers are unclear.
  4. Remove duplicate or outdated documents that may conflict with newer content.

Frequently asked questions (FAQs)​

Details

What AI model powers AI Employees? AI Employees run on enterprise‑grade LLMs that we evaluate and monitor for quality. Model names and versions are not disclosed and may change over time.

Details

Can I choose which model is used? AI Chat Receptionist and AI Voice Receptionist automatically use a supported model managed by Vendasta. AI Employees let you choose from the supported model options exposed in Business App settings, though we do not list the underlying provider names.

Details

Where do the assistants get their answers? From your configured capabilities, your knowledge base content, and the current conversation. They do not browse the open web beyond the sources you provide.

Details

How can I make answers more accurate? Provide clear, current content in the knowledge base and concise instructions in capabilities. Remove outdated or conflicting documents.

Details

Do AI Employees learn permanently from chats or calls? They use the active conversation for context. Persistent behavior comes from your capabilities and knowledge base, not from ad‑hoc user conversations.

Details

Can I prevent them from answering certain topics? Yes. Remove those topics from your knowledge base and add clear guidance in capability instructions about what the assistant should and should not answer.

Details

Do they use my website automatically? Only if you add your website as a source in the knowledge base. Otherwise, they rely on the documents and instructions you provide.

Details

Is my data shared with other businesses? No. Your account’s data, prompts, and responses are isolated. We monitor model performance in aggregate, but your content is not visible to other businesses.