Prompting Guidelines

Here are some important prompting guidelines to keep in mind when analysing your data using Petavue:

1. Be Specific and Clear: Use precise business terms rather than vague language.

   
   

   Example: Instead of saying "Show me customer retention related details", say "Show me customer retention rates by segment for Q4 2024"

   
   

2. Provide Context When Needed: Mention specific timeframes, customer segments, or business metrics. Include any relevant filters or conditions upfront.

   
   

   Example: "Analyze churn risk for enterprise customers who joined in the last 6 months with usage below 50%".

   
   

3. Ask Follow-up Questions Naturally: Petavue maintains context across follow-ups, so you can build on previous analyses.

   
   

   Example: After getting a churn analysis → "Now show me the common characteristics of the high-risk customers from that report".

   
   

4. Calculate Metrics on the Fly: When asking for undefined metrics, mention how the metric needs to be calculated.

   
   

   Example: “Calculate the Sales Cycle Length for the closed won opportunities for new businesses in the last quarter based on the Account created date and opportunity close date.”

   
   

5. Assign a Role to Claude: You can give Claude a Personality by asking it to act like a Data Analyst within your company and then ask it to fetch a report or perform an analysis.

   
   

   Example: Start the prompt with “You are a Data Analyst at Acme Inc. who is tasked to help me with my reporting and analysis request with the help of Petavue.

   
   

   Alternatively, you can do this as part of Instructions when creating a new Project within Claude.

To jumpstart your insights journey with ready-to-use prompts, refer to this link.