LinkedIn's Marketing Solutions blog published a playbook on 11 August 2026 titled "How B2B Marketers Can Dominate AI Search on LinkedIn 2026", and the number inside it that matters most is this: research from Meltwater cited in the playbook attributes 75 percent of LinkedIn citations in AI answers to individual member profiles rather than to company pages. The playbook wraps its guidance in a framework LinkedIn calls "The Credibility Stack", which is built on using the platform's existing credibility signals to push messages into AI generated answers.
A ratio of three to one against the company page inverts how most B2B teams are staffed. Budget, approval chains and content calendars sit with the brand account. If AI systems are overwhelmingly quoting people instead, the publishing asset a company most needs is one it does not own.
What LinkedIn actually published on 11 August 2026
The document LinkedIn released on 11 August 2026 is marketing guidance from the platform's own Marketing Solutions team, not a research report. It sets out The Credibility Stack as its organising idea, and it argues that the credibility signals LinkedIn already carries can be used to amplify a message across the platform and from there into AI answers.
Around that framework the playbook assembles figures from outside vendors and analysts. Profound, in a 2026 dataset, ranks LinkedIn as the most cited domain for professional queries in AI search. 6sense reports that 94 percent of B2B buyers use generative AI during research. Gartner predicts traditional search volume will fall by 50 percent by 2028. eMarketer, also in 2026, finds only 34 percent of B2B technology marketers feel prepared for an AI visibility strategy. LinkedIn discloses one number of its own, which is that its non-branded search traffic has declined by up to 60 percent.
The playbook also reports that B2B buying cycles compressed from 11.3 months in 2024 to 10.1 months in 2025, that articles account for roughly 60 percent of LinkedIn content citations against 40 percent for feed posts, and that Profound puts the median time for a new page to be cited at 6.81 days, with 90 percent of pages taking up to 37 days.
The 75 percent figure inverts the standard B2B publishing plan
Meltwater's 2026 finding, as cited in LinkedIn's 11 August playbook, is that individual member profiles produce 75 percent of LinkedIn citations while company pages account for the rest. Read against how B2B content teams actually operate, that is an awkward number.
The mechanism it implies is straightforward enough. AI systems appear to be treating a named person with a job title, a work history and a visible network as a more attributable source than a corporate account. The company page is an institution talking about itself. A member profile is a person with a record attached, and that record is exactly the kind of thing a system built to weigh credibility can read.
What follows for resourcing is less comfortable. If the company page is generating a quarter of the citations, then a content operation that publishes only through the brand account is working on the smaller share of the opportunity. Shifting that means the material has to go out under people's names, which changes who writes, who approves and who benefits.
The asset belongs to the employee, and that is a real exposure
This section is analysis rather than something LinkedIn states. The playbook recommends extending reach through employee profiles and relevant creators. It does not discuss what happens to that reach when an employee leaves.
A LinkedIn profile is personal property. The posts, the articles, the connections and whatever authority the profile has accumulated go with the person. A company that spends two years building citation share through four senior people has, at the end of it, four assets on someone else's balance sheet. When one of them moves to a competitor, the cited body of work moves too, and it keeps ranking.
There is no clean fix, but there are sensible hedges. Spread the effort across more people than the two most senior. Keep the company page publishing so the owned share is not zero. Republish or mirror the substance of an employee article somewhere the company controls, so the argument survives even if the profile does not. And be honest internally that personal brand building is part of the deal you are offering the employee, because that is what it is.
Articles out-cite feed posts 60 to 40, but the recommended cadence leans the other way
LinkedIn's 11 August playbook reports that articles produce roughly 60 percent of LinkedIn content citations while feed posts produce about 40 percent. Its recommended publishing rhythm is 2 to 3 posts weekly plus one article a week.
The format guidance is specific. Articles should run 800 to 1,200 words with clear headers and a TL;DR summary. Posts should run 200 to 300 words with keywords placed in the opening line. Content freshness is named as a priority.
Worth noticing that the cadence and the citation split point in slightly different directions. On a weekly basis the recommendation produces two or three short posts against a single long-form article, while the citation share favours the article. The reading that makes sense of both is that posts and articles are doing different jobs, with the feed carrying distribution and engagement and the article carrying the citable substance. LinkedIn does not spell that out, so treat the reconciliation as inference.
6.81 days is the number to put in front of an impatient stakeholder
Profound's figure, as cited by LinkedIn on 11 August 2026, is a median of 6.81 days for a new page to be cited, with 90 percent of pages taking up to 37 days. That is the single most useful line in the playbook for anyone managing expectations.
It sets a review window. Checking for citations two days after publishing tells you nothing, because the median has not arrived yet. Declaring a piece a failure at three weeks is also premature, since the distribution runs out to 37 days for the ninth decile. A programme that publishes weekly should expect its first readable signal somewhere around week five or six, once the earliest pieces have cleared the window and enough of them exist to show a pattern.
The figure also reframes what a bad month looks like. Silence in the first fortnight of a new publishing effort is the expected shape of the data, not evidence that the approach is wrong.
The numbers in the playbook, and who each one comes from
Because the playbook mixes LinkedIn's own disclosure with vendor and analyst research, attribution matters when any of this gets repeated in a deck. The table below keeps each figure attached to the party LinkedIn credits.
| Figure | Credited by LinkedIn to | What it measures |
|---|---|---|
| 75 percent of LinkedIn citations | Meltwater, 2026 | Share coming from member profiles rather than company pages |
| Median 6.81 days, 90 percent within 37 days | Profound | Time from a new page going live to being cited |
| 60 percent against 40 percent | LinkedIn's playbook | Citation split between articles and feed posts |
| 94 percent of B2B buyers | 6sense | Use of generative AI during research |
| Non-branded search traffic down by up to 60 percent | LinkedIn, its own data | Decline in traffic LinkedIn receives from non-branded search |
LinkedIn disclosing a 60 percent traffic decline is the most interesting line in the piece
LinkedIn stated in its 11 August 2026 playbook that its own non-branded search traffic has fallen by up to 60 percent. Platforms do not usually volunteer that kind of figure about themselves.
What follows is analysis, not a claim the source makes. A platform that has lost most of its non-branded search arrivals has a strong reason to want its users to believe the traffic is recoverable through a different channel, and to want publishing activity on the platform to increase rather than drift elsewhere. A playbook telling B2B marketers that LinkedIn is the top cited domain in AI search, and that the way to benefit is to publish more on LinkedIn, serves that interest cleanly.
That does not make the guidance wrong. Gartner's 50 percent forecast and 6sense's 94 percent figure describe a shift that is happening whether or not LinkedIn benefits from naming it. It does mean the piece should be read as a platform making the case for its own surface, with the incentive visible rather than hidden.
This is a platform's marketing content, not independent research
Saying this plainly is more useful than burying it. LinkedIn's Marketing Solutions team wrote the playbook to be persuasive, and every figure in it apart from the traffic decline comes from a third-party vendor or analyst whose underlying methodology is not published in the piece.
Meltwater's 75 percent, Profound's citation timings and its ranking of LinkedIn as the top cited domain, 6sense's 94 percent and eMarketer's 34 percent readiness figure are all cited without sample sizes, query sets or date ranges appearing in the playbook. The measurement of AI citations is a young discipline, and different vendors define a citation differently.
Treat the numbers as directional. Three to one in favour of personal profiles is worth acting on as a direction of travel. It is not precise enough to justify a formula, and a plan that only works if the ratio is exactly 75 to 25 is a plan built on a figure that has not been independently verified.
What this means for Thai marketers
The playbook contains no Thailand-specific data. LinkedIn reports no Thai sample, no Southeast Asia breakdown and no results for Thai language content, so anything below is reasoning from the global guidance rather than a claim the source makes about this market.
Thai B2B and professional services firms already use LinkedIn as a regional credibility layer, usually in English, and usually with a thin company page and a handful of active executives. Under Meltwater's 75 percent finding, that thin company page was never going to be the main asset anyway. The executives were. What changes is that their profiles stop being a networking convenience and become the publishing channel, which requires their time in a way a delegated company page never did.
The awkward part in a Thai organisation is usually hierarchy rather than capability. Asking a managing director to publish 800 to 1,200 word articles weekly under their own name, in English, is a real ask, and a ghostwritten piece that reads as though it came from the marketing department will not sound like a person. A workable version starts smaller, with one executive and a genuine subject they can defend in a meeting, and grows once the first pieces have cleared the 37 day window. Firms building a wider AI SEO programme can treat LinkedIn as one surface within it rather than a separate project, and pair the organic publishing with paid social distribution where reach needs help.
Common questions about LinkedIn's AI search playbook
Should we stop posting from our company page?
No, and the playbook does not suggest that. Meltwater's figure cited by LinkedIn puts 75 percent of citations with member profiles, which leaves a quarter with company pages. The reasonable response is to rebalance effort toward named people while keeping the company page active, not to abandon an asset the business actually owns.
Does any of this apply to Thailand or to Thai language content?
The source did not say. LinkedIn's 11 August 2026 playbook contains no Thailand-specific data, no Southeast Asian breakdown and no findings about Thai language posts. The guidance is written for a global B2B audience, and applying it here is an extrapolation.
How long before we can tell whether this is working?
Profound's numbers cited in the playbook give a median of 6.81 days to first citation and up to 37 days for 90 percent of pages. Set the first honest review at around five to six weeks after a weekly programme starts, which allows the earliest pieces to clear the window and gives you more than one data point to read.
Are these figures reliable?
Treat them as directional. The playbook is LinkedIn's own marketing content, and the figures come from Meltwater, Profound, 6sense, Gartner and eMarketer without sample sizes or methodology appearing in the piece. The one number LinkedIn reports about itself, the fall of up to 60 percent in non-branded search traffic, is the only first-party figure in the set.
What do we measure in the meantime?
LinkedIn splits measurement into three tiers, and the first is available immediately. Leading indicators are impressions and reactions, outcome metrics are citations and share of voice, and diagnostic patterns cover how different formats and different authors perform. Early on you will only have the leading indicators, which is why the diagnostic layer matters later for working out which people and which formats are carrying the result.
Where this leaves a B2B team
The practical shape of LinkedIn's advice is a weekly long-form article and a couple of short posts, published under named people rather than the brand, measured on a horizon of weeks rather than days. The unusual thing is a platform publishing the recipe for farming itself, at a moment when it has disclosed losing most of its non-branded search traffic. Both facts are worth holding at once.
If you want help deciding which of your people should be publishing, and what they should be publishing about, that sits inside how we approach content marketing. Talk to us and we will look at where your credibility currently lives.
