Skip to main content
1.6

The Tool Match

Estimated time: 25 minTool: Claude, Perplexity, Gemini
After this drill, you can:

After this drill, you have tested the same task across three different AI tools and can make a reasoned case for which tool fits which job.

Why this matters

The tool you use changes the output you get — not just in quality, but in kind. Perplexity cites sources. Gemini integrates with your Google Calendar. Grok knows what happened this morning. Claude reasons through complexity. Using the right tool for the job isn't about brand loyalty — it's about matching capability to need. This drill builds the habit of asking 'which tool fits this task?' before prompting.

How to do it

  1. 1

    Pick a research-oriented task that would benefit from cited sources

    Examples: "What are the latest findings on sleep and cognitive performance?", "What are the main approaches to reducing employee turnover?", "What has changed in EU data privacy law in 2024–2025?"

  2. 2

    Send the same prompt to Claude, Perplexity, and Gemini

    Use identical prompts for all three. Record each response.

  3. 3

    Evaluate each response on three dimensions

    Source quality (did it cite?), accuracy (did it match what you know or could verify?), usefulness (could you act on this?). Score each tool 1–3 on each dimension.

  4. 4

    Update your Tool Decision Guide

    Add or refine the "Research with cited sources" row based on what you observed.

The prompt

PROMPT — Research Task (send identical to all three)Model: Claude, Perplexity, and Gemini
[YOUR RESEARCH QUESTION]

I need this for: [brief context about why you need it]

Please cite your sources. I want to be able to verify the key claims.

Success criteria

  • ✓You tested the same task on three tools
  • ✓You evaluated each on source quality, accuracy, and usefulness
  • ✓You updated your Tool Decision Guide with at least one refined insight

Common mistakes

Using a task that doesn't benefit from source citations (creative writing, opinions)

→ This drill is specifically about research tasks. Pick a factual, current-events, or research question where source quality matters.

Not checking whether cited sources actually exist

→ Click one link per response. AI tools occasionally cite URLs that don't exist. The ability to spot this is a core AI literacy skill.