LinkedIn vs Substack: Where Does AI Find Your Expertise?
Updated: Oct 31
Writers give LinkedIn and Substack the same effort, and early 2026 data shows AI answer engines treat them as two different kinds of sources. One appears in answers constantly. The other barely registers.
That gap decides where your best thinking should live.
LinkedIn is a primary source for professional answers
Semrush studied 325,000 prompts across ChatGPT Search, Google AI Mode, and Perplexity in January and February 2026. LinkedIn ranked second among all cited domains in that sample, ahead of Wikipedia, YouTube, and every major news publisher. It appeared in 14.3% of ChatGPT Search answers, 13.5% of AI Mode answers, and 5.3% of Perplexity answers.
Profound tracked 1.4 million citations across six platforms from November 2025 through February 2026. LinkedIn climbed from roughly 11th to 5th among ChatGPT's cited domains in about three months, and ranked first for professional queries across all six platforms.
Treat those numbers as a snapshot of early 2026. The direction is the point: for questions about hiring, vendors, methods, and expertise, answer engines reach for LinkedIn.
People win on ChatGPT, brands win on Perplexity
Semrush also split cited LinkedIn URLs by who published them. On ChatGPT Search and Google AI Mode, 59% came from individual members. On Perplexity, 59% came from company pages.
A founder's profile and a company page feed different engines. Running only one leaves a platform uncovered.
Long-form beats the short post
In the Semrush data, LinkedIn articles make up roughly half to two thirds of cited LinkedIn URLs, depending on the platform. Short posts are the minority.
The studies measure what gets cited, not why. My read: an article is a complete page with a title, a named author, and a body that breaks into liftable passages. A post in a feed is one thought with nothing built around it. Engines pull passages, and articles contain more of them.
Substack builds readers, not citations
Britopian analyzed 1.7 million unbranded prompts and found Substack accounts for about 0.07% of all AI citations. Medium, in the same data, took 0.36%, more than five times the share.
Britopian sells AI visibility services, so hold the exact figure loosely. The pattern still matters. A newsletter with thousands of loyal paying readers can be close to invisible when ChatGPT, Perplexity, or Claude answer questions in its own topic.
The likely causes are structural. Every publication shares one root domain, so an individual writer inherits little authority of their own. Issues written for the inbox tend to open with story and bury the answer.
Paywalls can hide the useful part. None of this reflects writing quality. It reflects how the content is packaged and where it lives.
Some newsletters do earn citations, so the platform is not a dead end. It is a weak place to depend on.
What to do with LinkedIn
Publish articles, not only posts. Make each one answer a single question a buyer asks before they hire or buy.
Put the answer in the first two sentences. Name the method, the number, or the definition before any context.
Show your sources. Name where every statistic comes from and date the piece.
Cover both author types. Publish from your personal profile for ChatGPT and AI Mode. Run the company page as a real publishing channel for Perplexity.
Make the author obvious. A plain headline and About section that state what you are an authority on give each article a named expert behind it.
What to do with Substack
Treat it as a reader channel. It delivers your work to people who already know you.
Publish the full, answer-first version on a domain you own. Send the newsletter version after, with a link back.
Keep the answer on the free side of the paywall. Locked text cannot be quoted by a machine or a person who has not paid.
Open with the answer, not the anecdote. The story still has a place, just not before the claim.
Test it on your own questions
Write down ten questions a buyer asks before choosing you. Run each one through ChatGPT, Perplexity, Gemini, Claude, and Google AI Mode.
Record which pages get cited and where they live. Repeat every month. Citation patterns shift as models update, and these studies are a few months old.
The takeaway
The same insight can earn citations on one platform and disappear on another.
LinkedIn rewards a named author publishing a structured article. Substack rewards a writer with a real audience, and that audience does not carry over to answer engines.
Publish where machines look, and send readers to where they stay. Matching the format to the platform is the same fix I apply to content misalignment everywhere else: put the answer where the question gets asked.
Sources
Semrush, "We Analyzed 89K LinkedIn URLs Cited in AI Search" (March 10, 2026): https://www.semrush.com/blog/linkedin-ai-visibility-study/
The Keyword, coverage of Profound's LinkedIn citation research (2026): https://www.thekeyword.co/news/linkedin-ai-search-citations-professional-queries
Value Your Network, summary of Semrush data on LinkedIn article share (2026): https://www.valueyournetwork.com/en/ia-linkedin-2026-citations/
Britopian, "Substack and generative engines" (2026): https://www.britopian.com/geo/substack-generative-engines/
Comments