Stage 4: SYNTHESIZE (10 minutes)
Identify patterns, themes, debates, and gaps across all sources
From Data to Insight
Extraction gave you structured data—individual paper summaries with key findings. Synthesis reveals the big picture: what the literature as a whole says. This is where research becomes understanding.
The AI shows you patterns. You provide the judgment.
The 2-Step Synthesis Process
Step 4.1: Thematic Analysis (5 minutes)
Ask your AI assistant to analyze all the extracted papers and identify cross-cutting patterns using this synthesis prompt:
Prompt Template:
Based on all the papers we've processed, identify:
1. **Major Themes** - What are the 3-5 main topics or findings that appear repeatedly?
2. **Points of Consensus** - What do most papers agree on?
3. **Points of Debate** - Where do papers contradict each other or disagree?
4. **Research Gaps** - What questions remain unanswered?
5. **Methodological Patterns** - What approaches are most common? Any methodological limitations across studies?
Organize this as a synthesis framework I can use for writing.The AI will provide a structured analysis identifying recurring themes, areas of agreement, contradictions, gaps in the literature, and common methodological approaches or limitations.
Step 4.2: Your Critical Analysis (5 minutes)
The AI just gave you patterns—now you add judgment and contextual insight that only you can provide.
Ask yourself these critical questions. Which consensus points are actually well-supported by robust evidence versus just repeated claims without strong backing? Are the contradictions due to different methodological approaches, different study contexts or populations, or different time periods when data was collected? What gaps matter most for your specific research question and practical needs? Are there themes or patterns the AI missed that you noticed while reviewing the summaries?
Write 3-5 insights that represent your unique analytical contribution. These are the novel observations AI can't make because they require domain knowledge, practical experience, or contextual awareness.
Example Insights:
"Most studies focus narrowly on coding tasks with minimal attention to creative writing productivity, revealing a significant gap in understanding LLM impact across knowledge work domains."
"The 'productivity' measures vary wildly across studies—some measure task completion speed, others assess output quality, and still others track job satisfaction—making direct comparisons challenging and suggesting we're measuring different constructs under the same label."
"There's a consistent 6-12 month learning curve before productivity gains appear in empirical data, suggesting early studies conducted within months of AI tool adoption likely underestimate long-term impact."
Time Check: 45 minutes elapsed since Stage 1. You're now ready to write.
Why Human Analysis Matters
AI identifies surface patterns—co-occurrence, repetition, statistical trends across papers. But you understand causality, context, and significance. You know which findings matter in your domain, why contradictions exist, and what questions practitioners actually need answered.
Your 3-5 insights are what turn a literature review into an original research contribution.