How to Train Your Kaarya AI on Your Company's Data

A generic AI is like a smart intern who doesn't know your business. A 'Trained' AI is like your best employee who has been there for 10 years. To get the most out of Kaarya, you need to provide it with the 'Context' of your specific operations.

Three ways to train your Kaarya Agent

1. Document Ingestion (The Knowledge Base)

You can upload PDFs, Word docs, or link your website. Kaarya's RAG (Retrieval-Augmented Generation) engine parses these documents to answer customer questions about pricing, services, and policies accurately.

2. Strategic Prompting (The Personality)

This is where you define the 'Rules of Engagement.' How should the AI handle a refund request? What is the specific tone for your luxury real estate brand? You write the 'Instructions,' and the AI follows them perfectly.

3. Direct API Feedback (Live Data)

Kaarya can pull data in real-time from your CRM or Inventory system. It doesn't just 'know' your products; it knows if they are in stock right now.

Do	Don't	Why
Upload clean PDFs	Upload messy scans	Text clarity is key for AI
Use bullet points	Use long paragraphs	Clearer structure = Fewer errors
Update docs regularly	Set it and forget it	AI needs the latest info
Include FAQs	Assume AI knows everything	Specific answers drive conversion
Data Training Best Practices

Maintaining your AI's knowledge

As your business grows, your AI should grow with it. Kaarya provides a simple interface for you to add new documents or tweak instructions as your service offerings change.