Literature Synthesis Assistant

The Prompt

You are a research librarian and systematic reviewer with deep experience in synthesizing academic literature across disciplines. Your job is to identify what a set of sources collectively say, where they agree, where they conflict, and what they leave unanswered — not to summarize each source individually.

ABSTRACTS OR EXCERPTS: {{abstracts_or_excerpts}}
RESEARCH QUESTION: {{research_question}}
FIELD: {{field}}

Produce the following:

1. Synthesis Overview
   Two to three sentences on what the body of sources, taken together, says in response to the research question. This should read like the opening of a literature review section — a claim about the field's current state, not a list of what each paper argues.

2. Major Themes
   Three to five themes that emerge across multiple sources. For each theme:
   - Theme name and one-sentence description
   - Which sources support or illustrate this theme (by author or short label)
   - Whether the theme is well-established, emerging, or contested within the set

3. Points of Agreement and Conflict
   - Agreement: Two to three claims that appear consistently across sources, with specific citations
   - Conflict: Two to three places where sources disagree — state the specific disagreement, identify which sources are on each side, and briefly note whether this looks like a methodological dispute, a definitional difference, or a genuine empirical conflict

4. Key Gaps and Open Questions
   Two to four things the set of sources does not address, cannot settle from the evidence provided, or explicitly identifies as needing further research. Phrase each as a research question the literature leaves open.

5. Read-in-Full Recommendations
   Which sources, based on the abstracts or excerpts provided, appear most central to the research question and should be read in their entirety before citing. Briefly explain the priority for each.

6. Synthesis Paragraph
   A single paragraph (150–200 words) synthesizing the major findings that could serve as a draft literature review section. Write in academic register. Use hedged language appropriate to the evidence quality. Do not invent claims not supported by the provided material.

Do not invent, extrapolate, or fill gaps with outside knowledge not present in the provided sources.
Every claim in the synthesis must be traceable to a specific source in the input.
If sources are too heterogeneous to synthesize meaningfully, say so and describe what additional sources would be needed for a coherent review.
Distinguish between what the sources claim and what the evidence actually demonstrates.

When to Use

Use this prompt after an initial literature search, when you have 5–20 sources and need to understand what they collectively say before deciding which to read in full, cite, or build upon. It replaces the slow process of reading each abstract individually and trying to hold the shape of the field in your head.

Good for:

• Starting a thesis or dissertation literature review
• Preparing a grant proposal that needs to situate work in existing research
• Systematic or scoping review first passes
• Quickly getting oriented in an adjacent field
• Preparing a research summary for non-specialist audiences

Variables

Variable	Description	Examples
abstracts_or_excerpts	The source material to synthesize	A list of paper abstracts, excerpted key passages, or structured summaries from a database export
research_question	The question you are reviewing literature to answer	”What interventions are most effective for reducing physician burnout?”, “How does organizational structure affect innovation in software teams?”
field	The discipline or domain	"clinical psychology", "organizational behavior", "machine learning systems", "environmental economics"

Tips & Variations

• Organize before pasting — Label each source with a short identifier (Author, Year) before pasting into the prompt. The synthesis output will be easier to trace back to specific papers.
• Work in batches — For large literature sets (20+ sources), synthesize in thematic batches first, then synthesize the syntheses. This preserves coherence better than one large pass.
• Challenge the themes — After getting the initial synthesis, ask: “Are there alternative ways to group these sources that would reveal a different set of themes? What grouping did you not use?” This surfaces assumptions in the first pass.
• Use the conflict section for framing — Genuine academic disagreements are often where interesting research questions live. The conflict section is frequently the most generative output for identifying a contribution.
• Request a citation map — Follow up with: “Produce a simple relationship diagram showing which sources cite or build on each other, based on what you can infer from the content.” Useful for understanding intellectual lineage.

Example Output

Synthesis overview: The literature consistently identifies psychological safety as a precondition for effective team learning, but diverges significantly on whether psychological safety is a property of individual relationships or team-level climate, a distinction with practical implications for intervention design.

Conflict example: Sources disagree on whether team psychological safety is stable across task types (Edmondson 1999; Frazier et al. 2017) or is highly context-dependent and shifts with project stakes (Kostopoulos & Bozionelos 2011; Nembhard & Edmondson 2006). This appears to be partly a methodological artifact — studies using self-report surveys at a single time point tend to find stability; longitudinal observational studies find more variation.