HIVA

Platform Guide

Complete walkthrough for creating, training, deploying, and managing your AI health assistants.

1. Creating an AI Assistant

  • 1Navigate to the Assistants section in your HIVA Admin dashboard.
  • 2Click "Create New Assistant" and choose your assistant type (State Chatbot, Medichat, or Provider Support).
  • 3Configure the assistant name, description, welcome message, and target audience.
  • 4Select language preferences — HIVA supports English, Hausa, Yoruba, Igbo, and Pidgin.
  • 5Set response parameters including tone (clinical, friendly, formal) and response length.

2. Uploading Training Documents

  • 1Go to the Knowledge Base section of your assistant.
  • 2Upload PDFs, Word documents, or paste text content directly.
  • 3Supported content: health guidelines, drug formularies, clinical protocols, policy documents, FAQs.
  • 4The AI automatically indexes and learns from your uploaded content within minutes.
  • 5Review the training status and verify knowledge accuracy with test queries.

3. Configuring & Customising

  • 1Customise the chat widget appearance — colours, logo, position, and branding.
  • 2Set up conversation flows and fallback responses for unrecognised queries.
  • 3Configure access controls and user authentication if needed.
  • 4Enable multilingual auto-detection to serve users in their preferred language.
  • 5Set up notification rules for escalation to human operators.

4. Deploying Your Assistant

  • 1Generate an embed code snippet from the Deployment section.
  • 2Copy and paste the snippet into your website HTML — no coding required.
  • 3For mobile apps, use our REST API endpoints to integrate directly.
  • 4WhatsApp integration is available via our dedicated WhatsApp connector.
  • 5Test the deployment in preview mode before going live.

5. Monitoring & Analytics

  • 1Access real-time analytics from the Dashboard overview.
  • 2Track key metrics: total conversations, resolution rate, user satisfaction, response time.
  • 3Use the Chat with Your Database feature to ask natural-language questions about your data.
  • 4Export reports in CSV or PDF format for stakeholder presentations.
  • 5Set up automated weekly reports delivered to your email.

6. Chat with Your Database

  • 1Navigate to the Analytics → Chat with Data section.
  • 2Type natural-language questions like "How many beneficiaries enrolled this month?"
  • 3The system translates your question into a database query and returns results instantly.
  • 4Supports complex queries across enrollment, claims, and utilisation data.
  • 5Save frequently used queries as bookmarks for quick access.