Use caseWrite
Write a literature review from the papers you have already read
A review organised by theme rather than paper by paper — where the field agrees, where it does not and what nobody has tested — with every claim quoting the passage it rests on.
- You start with
- Twenty to sixty papers as PDFs in one folder, most of them read, some annotated. A question the review has to answer.
- Time
- An afternoon for the draft, after the reading.
- Your part
- The question, the themes, judging which disagreements matter, and checking the claims of consensus. Sorting papers into themes and the first draft are handed over.
What you'll end up with
- Sections by theme, each saying what the papers together show, not what each one says in turn.
- Disagreements between papers kept as disagreements, named, with both sides quoted.
- Footnotes that quote the paragraph each sentence relies on and open the paper at that passage.
A review that reads like a list of abstracts is the usual failure. The plan is where you stop that, before anything is written.
Most of the waiting is not spent working. It is spent waiting for someone to say yes…
- Everything above is ordinary Markdown in the file. Change a line and the draft follows it.
- Illustration, not a real run.
How to write a literature review from your papers with AI, step by step
Put the papers in one folder.
YouThe PDFs, and your notes on them if you have any. A scanned paper has no text until you run recognition on it, and until then it cannot be quoted.
Ask where they disagree before you plan anything.
You ask · MindMax answersAskWhere do these papers disagree about remote work and productivity, and is it about evidence, definitions or method?
This is the material a review is made of, and it is cheaper to find in a conversation than in a draft. Note the two or three disagreements the review cannot paper over.
Start a writing brief as a Literature review.
YouAskA literature review for chapter 2 of my thesis on remote work and productivity, organised by theme, from @papers. Show me the outline first.
Say it in the assistant, or put a writing brief in an empty document and pick Literature review from its genre menu. “Show me the outline first” makes it stop at the plan.
Rework the themes on the plan.
YouEach heading is a theme with a plan card under it: the point, how it is made, and the evidence — a named paper and the first sentence of the passage. Where papers conflict, the card has a Sources disagree line. Merge themes, move a paper, delete a card you do not want. This is the cheapest place to make it a synthesis.
Let it write, section by section.
MindMaxPress Write it all. Sections are written in order, the introduction and conclusion last, from what the body actually says. If the Sources line allows papers beyond your folder, they only fill gaps — and one known only by its abstract supports only what the abstract says.
Check every “most studies find”.
YouClaims of consensus are where reviews go wrong. Open their footnotes and check the quoted passages actually say it, and that the papers you know disagree are not missing from the sentence.
Read it through, then export.
YouIt reads the whole review once when it finishes; what it changed is listed on the brief and can be undone. Read it once yourself, then export to Word — each citation becomes a numbered source that keeps the quoted passage.
Why it works here
You see the themes before the prose
The plan names the papers and passages behind every section, as lines you can change — so the structure is your synthesis, not the order the files were read in.
AI writing →Quotes, not just a reference
Each footnote carries the exact words it relies on, so checking a citation is a click, not a trip back to the PDF.
Checked quotes →The papers answer questions first
Asking the folder where papers disagree reads the papers themselves, and every answer quotes the page it came from.
Ask questions across a folder →Also used here: AI PDF reader · Scanned PDF recognition · Export to Word and PDF · Guide: writing with a writing brief
Where it gets hard
- It reviews the papers you give it.
Coverage is your job. It synthesises the folder — plus, if the Sources line allows, a few papers to fill gaps — and does not tell you what the field has that your folder lacks.
- A theme can be an artefact of your folder.
Five papers by one group read like a consensus. Whether a pattern is the field’s or your reading list’s is a judgement only you can make.
- The citations are not APA or Chicago.
The export gives numbered sources with the quoted passage. Format the reference list in your reference manager or by hand to your department’s style.
- Checked quotes are not a fact-check.
A quote proves the paper says it, not that the paper is right. Weighing the quality of each study stays with you.