AI prompts for Research

This prompt breaks down the transcript of a quarterly results call or investor conference for equity research. It pulls out what management said, the analyst questions, financial and operating signals, warning signs, and less obvious hints, then writes an executive summary in bullets.

Produce a research questionnaire for tearing down a competitor.

This prompt summarizes a proposed bill or an existing law, lists the changes it would make, and contrasts how its backers present it with a balanced reading. It also points out parts that may get less attention, such as fees, loopholes, or cut programs, and their likely effects.

Identify logical fallacies and provide fairer restatements for an argumentative text.

This prompt tests one claim against a passage from a source in four fixed stages: pull out literal statements, rate the source's reliability, judge whether the claim is supported, then look for other readings. It ends with a verdict of supported, contradicted, or not enough info, plus a brief reason.

Plan phased hyperparameter experiments based on the model and available compute budget.

Summarize an invention article with its year, inventors, concepts, and description in JSON.

Extract central entities, descriptions, and relationships from a document as a knowledge graph in JSON.

Summarize a podcast transcript into key takeaways and quotable lines.

This prompt scores a study from 1 to 10 for novelty, rigor, and how empirical it is, and checks for conflicts of interest. It returns short bullet lists on the findings, methods, possible bias, and whether the work could be reproduced, then a letter grade.

Map the main arguments, agreements, and disputes in a Reddit thread.

Extract a research paper's title, authors, abstract, and keywords into JSON.

Extract citations, key quotes, and claim-evidence pairs from a research PDF paste.

Find the strongest claims in a sales deck and list what to fact-check.

Score positive, negative, and neutral sentiment, then assign a label and rationale in JSON.

Stress-test a startup idea for customer, wedge, and risks.

Define an agent role for statistical, correlation, and regression analysis with named tools.

Cluster open-ended survey responses into themes with JSON output.

This prompt writes a technical summary of a full research paper. It covers the title and authors, the main goal, the approach, what sets it apart, the experiments and results, strengths and limits, and the conclusion.

Classify text across five topics and return category scores with the highest-scoring label in JSON.