Make your work findable by machines
You publish as “B. Gorelik” in one paper, “Boris A. Gorelik” in another, “Gorelik B.” in a third. To an LLM-based search engine, that’s three people, and your work gets split across all three. Each of the three half-yous looks less accomplished than the real you.
Here’s the message, plain: if your paper doesn’t surface when ChatGPT, Perplexity, or Google AI Overviews are asked about your topic, your paper doesn’t exist for a growing fraction of your future citers. Profile hygiene is the cheapest, highest-value visibility action you’ve. It is unsexy. Do it anyway. It’s a one-afternoon job that pays for the rest of your career.
Why the machine cares about your profiles
LLM-based search engines build answers from publicly indexed text. They love structured author profiles (ORCID, Google Scholar, OpenAlex), plain-language summaries (a personal site, a LinkedIn About, an institutional-page paragraph), and consistent author-name disambiguation. They hate the three-spellings problem: they treat you as three people and undercount your work.
A clean profile lets recognition accumulate to you instead of to three confused half-yous. That is the whole game here. You are not optimizing an algorithm. You are giving the parser one consistent thing to attach credit to.
The checklist
Most of this is mechanical. Claim profiles, paste papers, fix the name spelling. Work down the list:
- Create or claim your ORCID iD. Add every paper. Pin your top 5.
- Claim your Google Scholar profile. Enable email alerts for new citations. Pin your top 5.
- Claim your OpenAlex author record. Verify the works list. Merge duplicate author IDs.
- Pick one canonical spelling of your name and use it on every future submission. Make ORCID, Scholar, and OpenAlex agree.
- Draft your 2-3 sentence plain-English bio (prompt below). Edit it. Add it to your institutional page.
- Mirror that bio word-for-word on your personal site, LinkedIn About, and the Notes field of your top ORCID and Scholar entries.
Spend two hours. You are done (approximately, update annually).
The one piece of actual writing
The whole tactic has exactly one part that’s writing rather than clicking: the 2-3 sentence bio for your institutional page, ORCID, and LinkedIn. Don’t write it from scratch. Draft it with the prompt below, edit until it sounds like a description and not a brag, then mirror it word-for-word across every page. The repetition is what makes a phrase canonical to an LLM.
You are a science communicator helping me write the 2-3 sentence "About" bio that appears on my institutional page, ORCID, and LinkedIn. The bio must be plain English, no jargon, and tell a smart non-specialist what I work on and why it matters in <SUB-SUB-FIELD>.
Hard constraints:
- 2 to 3 sentences. ~50 words total.
- First sentence: what I work on, in plain English.
- Second sentence: one specific result, dataset, or method I am known for, or building towards.
- Optional third sentence: why this matters to someone outside my sub-sub-field.
- No marketing language. No "passionate about", "expert in", "leading researcher", "groundbreaking". The bio should sound like a description, not a brag.
Inputs:
- My sub-sub-field: <PASTE>
- My top 3 papers (title + 1-line gist each): <PASTE>
- One sentence on what I'm working on now: <PASTE>
The LLM bio will be correct and slightly bland. The cheapest fix is to rewrite the second sentence in your own voice, naming the specific thing you’re known for with one concrete detail. Then mirror it word-for-word.
Audit it every quarter
Once a quarter, check whether the AI engines actually surface you. List 3-5 questions a smart non-specialist (a journalist, a grant officer, a PhD student outside your group) would type into ChatGPT or Perplexity about your topic. For each, ask the engine for the 5-10 most important researchers and papers to know. Then note: did your name appear? Your most-cited paper? A paper that should be there but isn’t?
If your work is missing, the diagnosis is one of four things:
- Profile-hygiene failure. Clean up ORCID, Scholar, OpenAlex.
- Plain-language failure. Write and post the plain-language summary.
- Conversation-presence failure. Your work hasn’t been discussed where LLMs ingest text, so write the answer that cites it.
- A genuine novelty gap. Too new. Write the summary and the discussion post, then wait 6-12 months.
This is a noisy signal. One bad run means little. Run it quarterly and watch the trend, not the single result.
Start with ORCID
If you do one thing today, open ORCID, claim your iD, and add every paper under one spelling of your name. That is the cheapest hour in this whole guide, and the rest of the checklist gets easier once the canonical record exists.
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.