Use this when you're choosing tools for a real task. It keeps the comparison tied to your files, permissions, and working conditions instead of a general popularity ranking.
Prompt template
Help me compare [candidate tools] for [specific task]. My audience, skills, and existing apps: [details]. Approved information: [details]. Required output and actions: [details]. Constraints: [quality, access, privacy, cost, and time]. Current official documentation: [links or excerpts; mark unavailable details as unknown]. Create a comparison table covering task fit, input/output formats, connections, permissions, account-specific data handling, limits, costs to verify, and operating effort. Design one small test with the same inputs and expected outcome for every candidate. Separate confirmed capabilities, vendor claims, your inferences, and unknowns. Do not invent current prices, rankings, integrations, or performance. Explain what evidence would justify a choice and what would make us reject it. Boundary: Produce only a comparison and a proposed test process. Keep supplied facts, requirements, and constraints intact, and leave source material unchanged. Don't connect accounts, purchase tools, or claim test results.
Fill in this prompt in the builder
What to supply
- Candidate tools, the task, your skills, and existing apps.
- Approved input, required outputs, and quality, privacy, budget, and time constraints.
- Current official documentation or excerpts; identify unavailable details.
Worked example
Illustrative example: the input and result below are invented.
Input
Lena, 31, wants a tool to turn an approved outline into editable slides. For this example, fictional Tool A's supplied documentation excerpt says 'exports editable slides; no offline mode.' Fictional Tool B's excerpt says 'exports PDF; works offline.' Both prices and retention rules are unknown.
Filled-in prompt
Show the complete prompt
Help me compare fictional Tool A and fictional Tool B for turning an approved outline into editable slides. My audience, skills, and existing apps: freelance designer; comfortable editing slides; existing slide editor. Approved information: an invented outline with no personal information. Required output and actions: editable slides; no sending or publishing. Constraints: editable output is required; budget and retention details are unresolved. Current official documentation: Supplied fictional excerpts: A exports editable slides and has no offline mode. B exports PDF and works offline. Prices and retention are unknown. Create a comparison table covering task fit, input/output formats, connections, permissions, account-specific data handling, limits, costs to verify, and operating effort. Design one small test with the same inputs and expected outcome for every candidate. Separate confirmed capabilities, vendor claims, your inferences, and unknowns. Do not invent current prices, rankings, integrations, or performance. Explain what evidence would justify a choice and what would make us reject it. Boundary: Produce only a comparison and a proposed test process. Keep supplied facts, requirements, and constraints intact, and leave source material unchanged. Don't connect accounts, purchase tools, or claim test results.
Result
A: Supplied documentation supports editable-slide export. Offline use is unavailable in the excerpt. Price and retention: unknown. B: Supplied documentation supports PDF export and offline use. Editable-slide export: unknown. Price and retention: unknown. Proposed test process: Use the same approved outline, inspect whether headings and slide text remain editable, and record correction effort. Decision evidence needed: Confirm A's retention rules and cost. Ask whether B has an editable format that the excerpt doesn't mention. Choice: Keep the decision open until the required output and account conditions are confirmed.
What I’d check or correct
I'd check that an undocumented feature wasn't treated as absent. Tool B's editable export is unknown, while Tool A's lack of offline mode is explicitly stated in the supplied excerpt.
What to check
- Capabilities are tied to current evidence for the relevant account and mode.
- The answer stays within comparison and a proposed test process, and leaves your requirements and supplied facts intact; it doesn't connect accounts or buy a tool.
- The same task and output criteria apply to each candidate.
Related pages
Sources
- AI Risk Management Framework · NIST · 2023. Checked October 8, 2026.
- Evaluation best practices · OpenAI. Checked October 8, 2026.