Loud Camel Loud Camel Notes on research visibility

The plain-language summary, and how to write one

By Boris Gorelik ·

Someone is going to ask an AI search tool what your paper says. The answer it gives back will be built from whatever plain text it can find about your work. If the only plain text is your abstract, the answer will read like your abstract: dense, hedged, and full of the word “moreover.”

So write the other version. A 200-word plain-English summary of every paper, posted alongside it, is the highest-value piece of writing you can do this year. It is the version of your work that LLMs will quote, reviewers will skim, and journalists will paraphrase. That is the whole message of this post.

Your abstract is not a plain-language summary

Your abstract is a structured 250-word document written for fellow specialists. It is full of jargon, full of hedging, and built to satisfy a reviewer in your sub-sub-field. That is a real job, and the abstract does it. It is just a different job from the one we are talking about.

A plain-language summary answers four questions in about four short paragraphs, around 200 words:

  1. What did we ask? One sentence, phrased so a smart non-specialist sees why anyone cares.
  2. What did we do? One or two sentences, method at a level your grandmother would tolerate.
  3. What did we find? Two or three sentences, the main result, with one number if you have one.
  4. So what? One sentence on why someone outside your sub-sub-field should care.

That is it. Four questions, four paragraphs, no “furthermore.”

Don’t write it from scratch

Writing this cold takes about thirty minutes, and most of that time goes to fighting your own instinct to sound like a journal. There is a faster way. Feed your abstract to an LLM with the prompt below, then edit the draft in five minutes.

You are an editor for a general-science magazine. You will receive the abstract of a peer-reviewed paper.

Rewrite the abstract as a 200-word plain-language summary, structured in four short paragraphs:
1. The question (1-2 sentences, phrased so a smart non-specialist understands why anyone would care).
2. The method (1-2 sentences, at a level a curious 16-year-old can follow).
3. The finding (2-3 sentences. If the abstract gives a quantitative result, include the number).
4. So what (1 sentence on why someone outside this sub-sub-field should care).

Hard constraints:
- No jargon. If a term is unavoidable, define it inline in one short clause.
- No words like "moreover", "furthermore", "interestingly", "novel". They add nothing.
- No hedging beyond what the abstract itself contains.
- 180-220 words total.
- Plain declarative sentences. Active voice when possible.

Abstract:
<PASTE ABSTRACT HERE>

The LLM produces the structure. You do the work that matters. Verify every claim against the abstract. Cut the jargon it missed. Add the one number that matters most. Rewrite the “so what” so it reads like you, not a model.

Read it adversarially, as a colleague in a different sub-discipline. If a sentence is not supported by the abstract, cut it. Cut anything that still sounds like jargon. The output is a draft, not a finished summary, and treating it as finished is the one way this goes wrong.

Put it everywhere

A summary that lives in one place does one job. The same summary, pasted in six places, does six. So paste it:

  • In the first paragraph of your personal site’s Publications page.
  • In the Notes field of your ORCID entry.
  • In the caption under figures you post.
  • In the first 200 words of a discussion post about the paper.
  • In the first paragraph of an email to people whose work you cited.
  • In the press release, if your university wants one.

Each of those is a surface where a reader, or a search tool, meets your work in plain text instead of jargon.

One more thing you can do with it

Once you have the summary, you have the raw material for a graphical abstract too. An LLM can draft a layout brief in two versions, one for peers and one for the general public, which you then render in Figma, BioRender, or Canva. That is a separate workflow with its own prompts, covered in the prompt-pack post. Mentioning it here only so you know the summary is the first domino.

Do this this week

Pick your most recent paper. Open the abstract, run the prompt, and spend five real minutes editing the draft against the four questions. Then paste the result into your Publications page and your ORCID Notes field today, before you close the tab. The next person who searches for your work will meet that version first.

This post is part of the Loud Camel field guide to academic visibility. You can read the whole guide as a single PDF: the handbook.