Mahal Systems CORGI
Fits business-analyst agents whose users ask why and what-next questions, not only lookups. Borrow its four-level question taxonomy (descriptive, explanatory, predictive, recommendational) and its success execution rate metric.
What it asks the agent to do
- answer a descriptive business question with SQL over a company-style database
- explain a trend by combining several queries
- make a prediction or forecast from the data
- give a recommendation that is grounded in query results
What to borrow from it
taxonomycase template
Caveats
No code or dataset link was found from the abstract page although the paper says one is released; databases are synthetic imitations of companies such as DoorDash and Airbnb; single-turn only; the paper is still under review and figures changed between versions.
Details
| Workloads | sql, data analysis |
|---|---|
| Interaction | single turn |
| Environment | database |
| End state checked | No |
| Tools | SQL execution |
| Grading | programmatic check, llm judge |
| Size | Not published |
| Data | synthetic |
| Personal data | none |
| License | unknown |
| Harness | none |
| Status | active |