The Ozgar AI chatbot brings a conversational layer to your documentation dashboard, letting you ask natural-language questions about any aspect of your application and receive context-aware answers in real time. Follow these guidelines and examples to get the most out of your AI assistant.

 

 

1) Getting Started

  • Open the Chat Panel Click the chat icon in the right-hand pane. The chatbot will automatically be aware of the current documentation page or visualization you’re viewing.
  • Frame Your Question Clearly Start with a “what,” “where,” or “how”—for example:
    • “Where is the order-validation logic implemented?”
    • “How does the payment workflow handle failures?”
    • “What are the possible states of the OrderStatus field?”

 

 

2) Best Practices

  • Be Specific Include object names, table names, or program identifiers to narrow the scope.
    • ❌ “How do I process orders?”
    • ✅ “How does ProcessOrderCL call UpdateInventoryDB?”
  • Use Follow-Up Questions The chatbot retains context. After asking “Show me the ER diagram for the Customers table,” you can follow up with “Which stored procedures update that table?” without restating the object.
  • Ask for Code Snippets or Diagrams Explicitly request examples if you need them:
    • “Can you show the RPG code snippet for the CalculateDiscounts routine?”
    • “Generate a flow diagram for creating an order”
  • Iterate to Drill Down Start broad (“What modules handle billing?”) and then drill deeper (“What lines in BillingService.java call the tax API?”).

 

 

3) Common Use Cases

Use Case

Sample Prompt

Locate Business Logic

“Where is customer credit-limit validation implemented?”

Trace Data Flows

“Which programs read from OrderHeader and write to OrderArchive?”

Understand Config

“What configuration parameters control log retention in the Order module?”

Troubleshoot Errors

“I’m seeing error code 5023 in production—what does it mean and how do I fix it?”

Generate Documentation

“Create a summary of the OrderProcessing service, including its API endpoints and database tables.”

 

 

4) Tips & Tricks

  • Use Wildcards If you’re not sure of the exact name, try partial matches: “Show me code for *Invoice* routines.”
  • Review and Refine If an answer is too broad, add qualifiers: “Only include calls that occur after the payment step.”

 

With these practices, the Ozgar AI chatbot becomes your personal guide — capable of surfacing hidden details, explaining complex flows, and even drafting code snippets. Whether you’re onboarding, debugging, or documenting, the chatbot accelerates your understanding of the entire application.