Why it is written this way
When asked to summarize literature, AI often defaults to sequential summaries: "Kim (2021) showed X. Lee (2019) showed Y." That produces an annotated bibliography, not a literature review. A literature review synthesizes how a field is divided and positions the gap where your own study belongs. The core purpose of this prompt is shifting the organizing axis from individual papers to overarching debate themes.
Allowing the model to clarify first in the second paragraph prevents premature assumptions. Notes often lack details like sample sizes or timeframes, and defining the right debate axes depends heavily on the researcher's specific focus. Limiting the model to a maximum of three questions ensures the workflow stays focused without descending into endless Q&A loops.
The format section sequences the table before the text. Generating the structural matrix first ensures the subsequent narrative paragraphs faithfully follow the structured evidence. The "Conflicting/Mixed Studies" column is crucial because where existing findings diverge is precisely where new research opportunities emerge.
The constraints prevent the AI from making editorial judgments on which study is "correct." When findings conflict, the review needs to identify methodological differences—such as study duration, sample scale, or measurement tools. Separating out an "Uncategorized Notes" section preserves thematic clarity without artificially forcing irrelevant studies into a debate. Wrapping the notes in """ delimiters prevents personal commentary within the notes from being misinterpreted as prompt instructions.
Unfamiliar terms? See Aha AI: context-window, hallucination
Compared with a bad example
Summarize my literature review notes.
(Paste paper notes here)
This outputs a passive, study-by-study summary. Even when studies directly contradict one another, their differences remain unanalyzed, typically ending with a vague cliché like "Various studies have been conducted." It fails to highlight research gaps, making it unusable as a draft.
Variations
When debate themes are not yet clear
Read the notes on "{{research topic}}" below and propose 3 to 5 candidate axes/themes for organizing these studies.
For each axis, provide a one-line explanation of "What this theme reveals" and list the study names that belong under it. Do not generate full tables or paragraphs yet.
Paper Notes: """ {{paper notes}} """
Use this variant when you are exploring multiple ways to structure your synthesis before committing to a final outline.
When searching for research gaps
Below are literature notes on "{{research topic}}". Identify the unresolved research questions that remain across these studies.
List them in this order: unstudied populations/durations/conditions, unresolved contradictions between findings, and incomparability due to different measurement tools. Include the cited notes in parentheses for each point. Do not suggest specific research topics.
Paper Notes: """ {{paper notes}} """
Useful for framing problem statements. Recommendation of new topics is explicitly restricted to keep the focus grounded in actual evidence.
Model notes
If you have more than 20 papers, split them into batches of 10 to generate separate synthesis tables first, then ask the model: "Merge the themes from these two tables into a unified matrix."
Related prompts
Last updated 2026-09-02 · Found a mistake? Let us know