Do not index
How do I get my content cited by AI chatbots? That question has replaced "how do I get more reach" in most of the founder conversations I have had this quarter. LinkedIn answered it publicly this month, and the answer is right about format and wrong about location.
Publish long-form. Do not publish it primarily on LinkedIn. That is the short version, and I will defend it.
LinkedIn released a playbook on optimizing content for AI visibility, built around a framework the company calls The Credibility Stack. Covering it for Social Media Today, Andrew Hutchinson noted that LinkedIn "shared insight into the specific content types that attract AI citations, noting that long-form articles in the app perform better than feed posts." Content freshness is flagged as a ranking consideration as well.
The format finding matches what I see. Depth gets cited. A 200 word feed post built on a hook and four line breaks does not give a language model enough to quote. A 1,200 word piece with a clear claim, a named idea, and specific numbers does. That part of the playbook is accurate and worth acting on this quarter.
The location advice is where it turns. LinkedIn telling you that AI citation favors long-form articles inside LinkedIn is LinkedIn telling you to move your archive onto its property. Hutchinson makes close to the same point in his coverage. You would be renting twice. Once from a platform that can change how articles surface whenever it wants, and once from AI systems that decide quarter by quarter which sources they weight. Neither landlord owes you notice.
Who this matters for and who should ignore it
This is for agency owners between $200k and $2M in revenue who already have a working content engine and are now deciding where the long-form version of their thinking should live. It is for ghostwriters charging $5k to $30k per month who are being asked by clients whether AI search changes the deliverable. It is for founders running personal-brand content who have published 200 posts and have nothing outside the feed to show for it.
This is not for anyone still trying to get their first consistent 90 days of publishing done. If you post twice a month and you are reading a playbook about AI citation, you are optimizing a distribution problem you do not have yet. Skip this if your bottleneck is output. Skip it also if you sell purely on referral and have no interest in being discoverable by strangers, because none of this changes your pipeline. And if you are still measuring content success by impressions, this article will not change your model, because the entire argument depends on caring about assets rather than reach.
The Archive Rule
Here is what I would actually do. I use what I call the Archive Rule, and it has one line. The long-form version lives where you own it, and the platform gets the excerpt.
In practice that means the full article publishes on your own domain first. The feed post goes up the same week, carrying the argument in compressed form, written to stand alone so it earns its own reach without depending on a click. The LinkedIn article, if you publish one at all, is a syndicated copy pointing back. You are not choosing between owned and rented. You are deciding which one holds the asset and which one does the distribution.
The reason this works is that AI systems weight source authority, and source authority accrues to a domain over time. A 3 person agency publishing 40 durable articles a year on its own site is building something that compounds across every model retrain. The same 40 articles inside LinkedIn build LinkedIn. That distinction sounds academic until the weighting shifts and 100 percent of your citable body of work sits behind someone else's login wall.
Freshness cuts the same way. LinkedIn flagging recency as a ranking factor means your archive needs maintenance, not just volume. Updating 12 strong articles a year beats publishing 60 that decay untouched. On your own domain you can revise a piece in place and keep the URL. Inside a feed you cannot, because the post is a moment rather than a document. Anyone thinking seriously about a durable LinkedIn content strategy runs into this the second they try to update something they published eight months ago.
The counter-argument I hear is that the audience is on LinkedIn and the domain gets no traffic. That is true and it is beside the point. The article on your domain is not fighting for feed attention. It is the thing a language model can reach without an account, the thing a prospect finds when they search your name after reading a post, and the thing that still exists if your account gets restricted on a Tuesday.
What this changes about your trajectory is narrower than it sounds. You are not going to see a spike. You are going to arrive at a point, probably 18 months out, where your name gets returned in an answer you did not know was being generated, sourced to a page you control. Operators who spent those 18 months filling someone else's archive will be starting that clock from zero, with a body of work that technically exists and functionally cannot be cited. The gap between those two positions is not talent or output. It is a decision about where the file lives, made early, and then not revisited.
