Research a big purchase without the review swamp
A shortlist of two or three options matched to your actual constraints, with the gotchas surfaced, in twenty minutes.
The problem
Buying anything over a few hundred dollars means two evenings lost to contradictory reviews and affiliate listicles.
What you get
A working system with the steps, tools, checkpoints, and expected return made explicit.
- Setup
- 20 minutes
- Back each week
- 1 hours
- Difficulty
- starter
Expected return
The working case
A planning estimate, not a guaranteed result. Measure the first four weeks against your own baseline.
48
hours returned per year
At 1 hours/week across 48 working weeks.
1
week to earn back setup
Compare the setup estimate with the weekly time returned.
Operating contract
Input
The source material, constraints, and examples a human would need to do this work well.
Checkpoint
A person reviews judgment calls, sensitive content, unfamiliar tools, and irreversible actions.
Success signal
Track time returned, corrections required, and exceptions. Keep it only if the measured result compounds.
The steps
- 01
Define the decision, not the product: what it must do, budget, dealbreakers.
Copy this prompt
Help me choose [a washing machine / laptop / stroller]. Constraints: budget [$X], must [fit 60cm / run silent / fold one-handed], dealbreakers: [none of brand Y]. Give me: 1) three candidates with honest tradeoffs, 2) the common failure points owners complain about for each, 3) three questions to ask or check before buying.
- 02
Ask it to steelman the cheapest option and the most expensive; the argument usually reveals what you actually value.
- 03
Verify current prices and stock yourself; treat model numbers and prices from AI as hypotheses, not facts.
What it runs on
- Perplexity
Search-grounded answers with sources; good for current models and prices.
- ChatGPT
With web search on, comparable; better at the tradeoff reasoning.
Where this goes wrong
- AI invents model numbers and misremembers specs; always click through to the manufacturer page before paying.
- Recency matters for electronics; ask it to restrict to models from the last 18 months.
Definition of done
Run it for four weeks. Then make it earn its place.
- □ Baseline the manual time before launch.
- □ Keep a human approval step for consequential output.
- □ Record corrections and exceptions, not just successes.
- □ Expand, revise, or retire it after the first review.
Build the system around it
Related workflows
If this one stops working, tell us. Three reports in a month and it leaves the library until a person has looked at it again.