Skip to main content
Question

AI Agent mission and month variable in scenarios not using correct month and year

  • November 22, 2025
  • 0 replies
  • 2 views

I need help fixing a mission for a variance analysis agent.  I am attempting to make it as dynamic as possible, so I am using a variable called @CurrentMonth which is a calculated metric that pulls the most current month we have transaction data into all my scenarios.   The calculated value for the @CurrentMonth variable for my @Actuals scenario is “0ct 25”.  I have another scenario called “Working Forecast” that has a calculated value of “May 25”.

When I run this AI agent, no matter how I write the instructions, it uses the value of “May 25” which is nowhere in the instructions, not even implicity.

The only way this works is if I change the variable for @CurrentMonth to be hard coded as “Oct 25”.

What am I doing wrong? How would you change it?

Mission instructions below,

Analyze @Revenue and @Operating Expenses metrics, filtering for:

  • Reporting Time = "MTD"

  • @CurrentMonth  = @Actuals scenario

Structure the report in these sections:

  1. Conclusion: Summarize the main findings.

  2. Key Highlights: List key points for @Revenue and @OperatingExpenses 

For all analysis:

  • Compare @Actuals scenario to @Budget scenario at the @GLAccountLvl0 level.

  • Only show data for @Actuals and @Budget side by side

  • Break down by @BudgetArea.

  • Filter to show only @BudgetArea with a negative variance of at least $10,000 and at least 5% negative difference.

  • Include both % and $ variance in the output.

Recommendations: Provide concise, actionable recommendations based on the findings.

Formatting and Style:

  • Use simple, clear language.

  • Keep the report concise.

  • Use bullet points or tables for highlights and variances.