Hour 1 · SLR · Step 7 of 1547%
1.5
Synthesis & write-up
From a clean extraction table to a defensible narrative.
~10 min
Synthesis turns rows into patterns. AI is good at first-draft thematic groupings and at smoothing methods prose; it is bad at deciding what matters. You decide that.
AI prompts (2)
Prompt
Thematic synthesis (first pass)
When: You have an extraction table and want candidate themes to react to.
You are helping me thematically synthesise a set of studies. Extraction table (CSV or markdown): <PASTE> Please: 1. Propose 3–6 candidate themes that cut across the studies. For each, give: theme name, one-sentence definition, the study IDs that belong to it, and one representative quote or finding. 2. Note any study that does not fit any theme — do not force-fit. 3. List 2–3 tensions or contradictions across the corpus. 4. Suggest 2–3 research gaps the corpus implies. Be honest about uncertainty. If a theme is thin (≤2 studies), say so.
Prompt
Methods section drafter
When: You have your PRISMA numbers and want a clean methods paragraph.
Draft the Methods section of a systematic literature review manuscript using the inputs below. Databases searched: <list + dates> Search string(s): <paste> Inclusion criteria: <paste> Exclusion criteria: <paste> Screening tool used: <e.g. Rayyan> Number of reviewers and conflict resolution: <e.g. two reviewers, third for conflicts> PRISMA numbers: identified=<n>, after duplicates=<n>, screened=<n>, full-text assessed=<n>, included=<n> Quality appraisal tool: <e.g. MMAT, JBI> Write in formal academic English, third person, past tense. Cite PRISMA 2020 (Page et al., 2021). Do not invent any number or step not listed above.