c_Learn & Practice AILEARN/Build a useful operations dashboard
← Cookbook
COOKBOOK

Build a useful operations dashboard

Create a dashboard from {{SAMPLE_DATA}} for {{AUDIENCE}}. Define each metric, date range, units, and missing-data behavior. Label sample data clearly.

Define the problem

Create a dashboard from {{SAMPLE_DATA}} for {{AUDIENCE}}. Define each metric, date range, units, and missing-data behavior. Label sample data clearly.

Inspect prerequisites

Read project instructions and relevant configuration. Record the current state, use sample data, and confirm a recovery point.

Make a bounded plan

Identify the smallest change, its affected files, and a testable expected result. Keep publication separate.

Implement in small steps

Use the linked prompt with real project details. Review each stage before extending it. Preserve unrelated work.

Verify and debug

Recalculate totals independently and test empty, zero, and invalid records.

Review and hand off

Inspect the diff and document changes, passed/failed/unrun checks, remaining limits, and recovery. Ask the owner to approve any live deployment.

Prompt library

GOAL:
Create a dashboard from {{SAMPLE_DATA}} for {{AUDIENCE}}. Define each metric, date range, units, and missing-data behavior. Label sample data clearly.

CONTEXT / INPUT:
Project: {{PROJECT_NAME}}. Read relevant files and project instructions first. Ask only about material missing information.

SCOPE:
{{SCOPE}}

DO NOT:
Do not overwrite unrelated work, expose secrets, delete data, publish, or deploy. Respect {{CONSTRAINT}}.

EXPECTED OUTPUT:
A reviewable result with changed files or findings, and a concise explanation.

ACCEPTANCE CRITERIA:
{{SUCCESS_CRITERIA}}

VERIFY:
Recalculate totals independently and test empty, zero, and invalid records.

REPORT:
Report changes, reasons, passed/failed/unrun checks, remaining limitations, and recovery steps. Never claim a check passed without evidence.

Common mistakes

Missing inputs, incorrect working folders, and unverified commands are common failure points. Stop at the failing step and gather evidence before trying again.

Before you run

Real projects
Esc · Close