Chat Assistant
The Arif AI chat assistant is a conversational interface that lets you interact with your Mjara ERP system using natural language. It appears as a floating widget in the bottom-right corner of your screen.
Features
Natural Language Interaction
Simply type your request in plain language. Arif AI understands context, remembers conversation history, and can execute operations on your behalf.
Example prompts:
- "List my recent sales invoices"
- "Create a new customer named ABC Trading"
- "What is my total revenue this month?"
- "Help me set up my company"
Real-Time Streaming
Responses stream token-by-token via Server-Sent Events (SSE), providing a natural conversational feel with minimal perceived latency.
Tool Execution
When the AI determines that a tool is needed, it automatically invokes the appropriate tool and shows the execution status. You can see which tools were called and their results directly in the chat.
Quick Actions
The chat widget provides quick action buttons for common tasks:
- Import Data — Upload files or records
- Create Invoice — Generate a new invoice
- System Setup — Configure settings
- Get Support — Access help and support
Conversation Management
- New Conversation — Start a fresh chat session
- Search — Search through previous conversations
- History — Access past conversation threads
- Archive — Archive old conversations
Message Versioning
You can edit a previous message to regenerate the AI's response. Each edit creates a new version, and the full version history is tracked for audit purposes.
Dark & Light Mode
The chat widget supports both dark and light themes to match your workspace preference.
Chat Flow
Multimodal Support
Arif AI supports multimodal conversations — you can attach images and documents to your messages for the AI to analyze.
Supported file types include PDF, Word documents, Excel spreadsheets, and images (with OCR support).
Context Auto-Summarization
For long conversations, Arif AI automatically summarizes earlier messages to maintain context without exceeding token limits. This ensures seamless, coherent conversations even across extended sessions.