ChatGPT agent analyses six Indian budget years

Vasu Deo Sankrityayan tests ChatGPT agent on India's Union Finance Budgets from 2020 through 2025. The brief specifies sector-wise allocations, labelled currency units, an agriculture trend and a comparison of absolute and percentage changes between the first and last years.

The author reports watching the agent consult government budget reports for all six years. It then creates a spreadsheet, a summary table, an agriculture-allocation plot and a sector-comparison bar graph. The original article links the spreadsheet, making the completed artifact more concrete than a claim that the model merely answered a budget question.

The reported eighteen-minute run is the author's observation, not a controlled timing benchmark. The review praises government sources but does not provide an independent row-by-row validation of the extracted figures, unit consistency or cross-year sector definitions. The useful brief therefore names the years, output columns, units and desired calculations before the agent collects data. A reviewer still needs to inspect whether a numerical change reflects comparable allocations rather than a renamed or reorganised sector. The original prompt excerpt below begins the actual tested instruction; the full source also defines the chart and summary requirements.

Prompt
Original prompt

Source: Blog · Vasu Deo Sankrityayan