Skip to main content
DocumentationLibraryReference
Cookbook
Prompt template
Everyone

Compare with a benchmark

Each value of a dimension beside a benchmark value that you select.

Group
Compare in context
Blanks
Semantic model, Measure, Dimension, Benchmark
Kater SQL
EVAL(measure, SET(dimension = :benchmark))

Prompt template

Fill in the blanks

Prompt
Use my Kater data to answer this question: How does each compare with on ? --- Data: Measure: Use the measure as Kater defines it. Do not write its calculation again. Show for each , the benchmark value beside it, and the difference. Get the benchmark value from the measure with set to . Make the benchmark a parameter, with as its default. Write a Kater SQL query, then run it. Show me the SQL, the result and one sentence on what stands out. If a name I gave does not match a field, show me the closest fields and ask me before you continue.

Fill in the blanks, then copy the prompt into the Ask panel in Kater or into your own agent (Connect your agent). A blank you leave empty stays in brackets, and your agent asks you for it.

What your agent does

  • Finds the semantic model, the measure and each other field you named. When a name does not match, it shows the closest fields and asks you.
  • Gets the benchmark from the measure with the dimension set to the benchmark, so every row compares with the same value.
  • Runs the query and shows you the SQL, the result and one sentence on what stands out.

Example

For the Ad Reporting semantic model (ad_reporting), the question is:

How does each Platform compare with Google Ads on Spend?

Select Use the example to fill in every blank with it.