1. AI Technology Used
WebOrigin may use one or more third-party AI providers, including Google Gemini and other AI services, to generate AI-assisted responses. The specific provider may change as we expand our AI infrastructure.
What this means for your data:
- Conversation text is sent to our AI providers (e.g., Google's Gemini API) to generate AI responses
- Providers process this data under their respective Privacy Policies and API terms
- WebOrigin does not store intermediate AI inference data beyond what is needed for the platform
- We have a provider abstraction layer — no single AI provider is permanently locked in
2. What the AI Does
✓Reads end user WhatsApp messages and searches your Knowledge Base for relevant answers
✓Generates natural language responses based on your uploaded training content
✓Captures lead information (name, phone, course interest) from conversations
✓Classifies conversation intent to route to the correct AI agent
✓Detects when a human counselor takeover may be needed
3. What the AI Does NOT Do
✗It does not access the internet or external websites during a conversation
✗It does not access other institutes' data or Knowledge Bases
✗It does not make financial, legal, or medical decisions
✗It does not guarantee 100% accuracy of responses
✗It does not have persistent memory of conversations across sessions (unless logging is enabled)
✗It does not generate content from outside your uploaded Knowledge Base
4. Your Responsibilities with AI
AI accuracy depends entirely on the quality and completeness of your uploaded training content. You are responsible for the AI responses sent to your end users.
- Train the AI with accurate content — Keep your Knowledge Base up to date with correct fee structures, schedules, and policies
- Test before going live — Use the AI Simulator to verify AI responses before activating for end users
- Monitor flagged responses — Regularly review AI messages flagged by end users or counselors
- Do not rely on AI for legally sensitive information — Always have humans handle financial agreements, legal commitments, and medical guidance
- Keep content copyright-compliant — Only upload content you have the right to use
5. AI Content Ownership
- AI responses generated on your account using your training content belong to you
- WebOrigin does not claim ownership of AI-generated content
- You are responsible for AI responses sent to your end users under your organization's name
- Your Knowledge Base content is never used to train AI models for other organizations
6. AI Transparency
We believe end users have a right to know they may be interacting with AI:
- We recommend including a disclosure in your WhatsApp business profile description: "Powered by AI — reply HUMAN to speak with a counselor"
- The human takeover feature allows immediate switch to human when an end user requests it
- WebOrigin does not facilitate deceptive AI impersonation of named human individuals
Contact for AI Policy Questions: contact.weborigin@gmail.com · WebOrigin · MSME: UDYAM-DL-02-0118375