Ku is your built-in AI assistant that understands your business data. Ask questions, give instructions, and Ku will use the right tools to get things done — from creating journal entries to checking inventory to managing schedules.
Requirements:
- AI tokens are required for each conversation.
- The tools available to you depend on your role permissions and which modules are enabled.
Getting Started
- Click the AI chat bubble in the bottom-right corner of any page, or go to AI Assistant in the sidebar for the full-page view.
- Type your question or instruction in natural language.
- Ku will respond, ask clarifying questions if needed, and use tools to take action.
- For actions that change data (creating expenses, journal entries, etc.), Ku will show a preview and ask for your confirmation before proceeding.
What Ku Can Do
Ku has access to tools across three domains, based on your permissions:
Accounting
- Create, search, and void journal entries
- Record expenses and supplier bill payments
- Look up account balances and chart of accounts
- Create suppliers and manage supplier data
- Pull financial summaries and GL reports
Inventory
- Check stock levels and product details
- Create purchase orders
- Search products and categories
- View supplier information
HR & Staff
- Look up employee schedules and attendance
- View attendance summaries and reports
- Assign missions and routine tasks
- Send messages to team members via Telegram
- Search employee information
Image Understanding
You can attach images to your messages (receipts, invoices, documents). Ku will automatically switch to a vision-capable model to read and process the image content.
Tips
- Be specific: "Create a journal entry debiting Utilities Expense for 5,000 and crediting Cash" works better than "record an expense".
- Ku remembers the current conversation, so you can ask follow-up questions.
- Each "New Chat" starts fresh — previous conversations are not carried over.
- For bulk receipt processing, the Receipt Processor is faster and uses 30-40x fewer tokens than the chat-based approach.