Rewrite a Research Prompt
'Research X and tell me what you find' returns a topic summary, not an answer. The rewrite adds the control a research ask needs: scope, criteria, and output shape.
Overview
Weak research prompts are unfalsifiable — 'find the best option', 'be thorough' — so the model returns a confident overview of everything and a recommendation of nothing. The rewrite makes the ask answerable: 'be thorough' becomes a coverage list you name, and the missing format, length, and exclusion lines become slots, so the result arrives as a decision input instead of an essay. This resource loads an open-ended research ask and applies the control-focused rewrite.
Workflow
-
Rewrite with Add Output Control
'Be thorough and detailed' becomes a coverage slot; 'a good one' becomes success criteria; format, length, and exclusion lines appear.
-
Name the coverage
List the 3–5 things the research must evaluate — pricing at your size, deliverability, integrations. That list is the question.
-
Set the output shape
'Comparison table, then a one-paragraph recommendation' turns the result from reading material into a decision input.
Why This Works
- Named coverage criteria give the model something to be wrong about — that's what makes research checkable
- An output shape forces ranking and trade-offs instead of even-handed summary
- Exclusions ('no vendor marketing claims') cut the lowest-quality source of research filler
Best for
- Decision-support research with a real choice at the end
- Anyone whose research prompts return Wikipedia-shaped overviews
- Recurring vendor/tool evaluations
Not for
- Comparing an open vs scoped research prompt side by side — that's the Prompt Comparator
- Reviewing how a research prompt changed across edits — that's the Prompt Version Diff
Use cases
- Sharpening a research ask before spending a long model run
- Converting 'tell me what you find' into a decision-ready output spec
- Teaching the difference between a topic summary and an answer