Capstone: design a one-season agri-AI pilot for your district

Three ways to see it

1. 1

   Section one: choose and justify. Name the district, the crop or animal, and the decision the AI will improve. Cite one local statistic that motivates the choice such as average yield gap or known pest outbreak frequency. State the baseline: how decisions are made today and why they fail. A tight problem statement makes everything that follows easier.

2. 2

   Section four: governance, ethics, and consent. PDPA 2023 alignment for any personal data. Farmer consent flow in the local language. Data sharing agreement with the partner institution. Equity check that the tool does not exclude smallholders, women, or remote farmers. Independent advisor on the project who is not paid by you and can flag problems. Document each of these before week one.

3. 3

   Section five: metrics, timeline, and exit. Three primary metrics tied to farmer outcomes such as yield, cost, or revenue. Three secondary metrics for equity and safety. A sixteen to twenty-four week timeline with at least one mid-season review. A pre-agreed exit rule: which metric below which threshold pauses or stops the pilot. Plan a closing event with farmers, partners, and the local department to share findings.

The why-tree

Try this with Claude

Submission and next step. Ten pages or less. Share with one farmer, one extension officer, one academic, and one regulator if you have the access. Ask each for the weakest section and rewrite it. Then secure funding for the pilot. Pakistani agriculture needs careful, named pilots more than it needs new platforms. Your capstone is your ticket into that work.