Skip to main content

​
Effective agent interaction

Start your conversation by confirming your agent understands its purpose and testing its basic capabilities with simple requests.
Begin with straightforward tasks before moving to complex ones to understand your agent’s capabilities.

​
Key interface features

  • Message input for sending requests
  • File attachment support for document analysis
  • Conversation memory that maintains context
  • Controls for managing chat history and sharing

​
Prompting techniques

Be specific

Instead of: “Give me some ideas”
Try: “Give me 5 blog post ideas about sustainable gardening for beginners”
Example: “I’m preparing for a content strategist interview. Can you help me prepare answers for questions about analytics and SEO?”
Example: “Let’s analyze this data step by step:
  1. Identify top 3 performing products
  2. Suggest reasons for their success
  3. Recommend strategies for lower performers”
Refine responses by asking clarifying questions.
Example: “Could you make these instructions more concise and bullet-pointed?”

​
Advanced features

File analysis

Upload documents (PDF, DOCX, TXT, CSV) for analysis.After upload: “Summarize the key points and highlight unusual trends.”

Memory management

To manage conversation context:
  • Start new threads for different topics
  • Clear history to reset context
  • Reference specific parts of previous conversation
Always verify critical information from trusted sources, even when your agent is designed with specific knowledge domains.

​
Troubleshooting

If you’re not getting desired responses:
  1. Use more specific language
  2. Provide examples of what you’re looking for
  3. Simplify complex requests
  4. Ask about the agent’s knowledge of the topic
  5. Restart the conversation if needed
Consider refining your agent’s description or using the Code Editor if it consistently struggles with certain tasks.

​
Next steps

Share Your Agent

Learn how to share your agent with teammates, friends, or the public