With Analyst copilot, you can use natural language prompts to query your business information. In this article, you'll find tips, tricks, and examples to help you get started.
This article contains the following topics:
Writing effective prompts
These tips can help you write better prompts to get the most from Analyst copilot:
- Outcome: Be precise about what you want. For example, don't use "Show me tickets from last week." Try "In a table, show me the ten agents who solved the most tickets last week."
- Vagueness: Don't be vague. State exactly what you want. For example, don't use "Am I solving more tickets this year?", try "Show me a comparison of this week's ticket counts compared to the same week last year."
Example prompts
In this section, you'll find some example natural-language prompts you could use to build Analyst copilot reports.
Understanding staffing needs
- "For the last 30 days, compare ticket volume by channel (voice vs messaging vs tickets)."
- "Show ticket reopens in the last 24 hours by channel."
- "Show ticket median first reply time by month (hours)."
- "Why are my ticket numbers so high?"
Optimizing channels
- "Show a donut chart of tickets by channel in 2025."
- "How many times do customers make repeated contact, as new tickets using multiple channels within 30 minute windows? Segment this based on time of day."
- "On a day-by-day basis, which of my channels is the busiest?"
Checking compliance
- "How many tickets have the tag "Finance?"
- "Which tickets have breached their SLA this week?"
- "Show me all tickets with a first reply time greater than an hour."
Managing agent performance
- "Who didn't solve any tickets today?"
- "Show me the top ten ticket solvers last week."
- "Create a table showing, in order, the ten agents with the most tickets that have breached their SLA. Don't show John Smith or Jane Doe in the table."
- "What changed in mid-July, that drove my resolution times down?"