Research & comparison
The failure mode of research with a language model is not that it refuses to help. It is that it helps too smoothly: a confident paragraph, a plausible number, a citation that does not exist. Every prompt below is written against that. They separate what a source actually says from what the model inferred, ask for uncertainty to be marked rather than smoothed away, and force a recommendation instead of a tidy balanced list. Use them when you are comparing tools, mapping an unfamiliar field, checking a statistic that sounds too good, or turning a vague question into something you can actually go and answer.
36 prompts in 5 groups
Research & comparison21 categories
Starting on a new topic
Mapping an unfamiliar field, scoping the question, and finding somewhere real to begin reading.
- Map a field I am new toLongWeek one in an unfamiliar domain, before you can tell signal from noise.
- Turn a vague question into researchable onesLongThe question is important, and too big to start on.
- Five sources, none of them inventedMediumStarting on a topic and needing somewhere real to begin reading.
- Brief me on this topicMediumWalking into a conversation about something you do not know.
Reading and synthesis
Getting the substance out of a document, and reconciling sources that do not quite agree.
- Summarise this long document faithfullyLongA report, contract, or spec you have to understand but not read twice.
- Read this paper with meLongA research paper whose abstract promises more than you suspect it delivers.
- What is this source not saying?LongA press release, vendor page, or announcement that reads suspiciously well.
- Synthesise several sources into one viewLongYou have read five things on a topic and they do not quite agree.
- Write a briefing note from my researchLongTurning a pile of notes into one page a busy person will actually read.
- Summarise this in three pointsShortA long piece you need the substance of, quickly.
- What is this source quietly avoiding?ShortReading something that feels a little too tidy.
- Two sources that disagreeMediumTwo credible pieces reaching opposite conclusions.
Testing a claim
Statistics, causal stories, methodology, and tracing a repeated claim back to where it started.
- Extract the claims and rate the evidenceLongBefore you cite something persuasive back to your team.
- Is this causation or just correlation?LongAny claim of the form "X leads to Y" that someone wants to act on.
- Find the methodology weaknessesLongA study, survey, or internal analysis whose conclusion you are about to rely on.
- Sanity-check a market sizeLongA TAM slide, a forecast, or any big number in a deck you have to defend.
- Is this statistic plausible?LongA number that would be remarkable if true, which is exactly why it is being quoted.
- Trace this claim back to its originLongA "studies show" claim that everyone repeats and nobody links.
- Is this number plausible?ShortA statistic in a deck or an article that you cannot verify.
- Which claim here is weakest?ShortChecking an argument before you repeat it to someone else.
- Correlation or cause?ShortA headline implying that one thing caused another.
- Take this claim apartMediumA single claim you need to know whether to trust.
Comparing options
Weighing real candidates against your criteria, from a quick side-by-side to a full vendor evaluation.
- Compare these options against my criteriaLongTwo or more real candidates and a decision you keep postponing.
- Build a weighted decision matrixLongWhen the comparison has enough moving parts that you cannot hold it in your head.
- Evaluate a vendor or tool properlyLongBefore a trial, a contract, or a migration you cannot easily undo.
- Competitive analysis that is not a feature tableLongWhen you need to understand a rival's strategy, not just their pricing page.
- Compare these two, brieflyShortTwo options and no time for a full matrix.
- Compare these for my situationMediumA decision where generic comparisons have stopped helping.
Pressure-testing a conclusion
Counter-evidence, gaps, and the strongest objection, found before someone else finds it for you.
- Find the counter-evidenceLongWhen the research so far has all pointed the same way and that makes you nervous.
- Steel-man the opposing viewLongBefore arguing against a position you already think is wrong.
- What is the strongest objection to my conclusion?LongBefore you present a finding to people who will push back on it.
- Strongest objection to thisShortBefore you commit to a conclusion you already like.
- What would change this conclusion?ShortDeciding how firmly to hold a view you have formed.
- Who disagrees with this?ShortFinding the other side before someone else finds it for you.
- What have I not looked for yet?ShortThe end of a research session, before you write anything up.
- Stress-test my conclusionMediumBefore you put a conclusion in front of other people.
Got one that isn't here?
Send a prompt you keep coming back to. A person reads every one.
Suggest a prompt