LinkedIn posts now rank on Google and shape AI answers

LinkedIn's feed was never meant to be a search index. It has become one anyway.

Sendible Insights documents a striking demonstration of how fast this loop now closes: a strategist posted a sardonic, self-declared claim on LinkedIn, and within hours Google's AI Overview was citing it as evidence of the fact. The post had not been peer-reviewed, endorsed, or linked to from anywhere. It had been published, indexed, and absorbed.

That is not a curiosity. It is a distribution mechanism that most B2B communications teams have not yet priced into their editorial strategy.

The indexing gap most brands are ignoring

Google has long crawled LinkedIn's public posts. What has changed is the speed and the downstream consequence. A post that earns enough early engagement now appears in Google Search results, sometimes outranking the author's own website. The same post feeds large language models that use LinkedIn as a training corpus, and those models shape AI Overviews, ChatGPT answers, and Perplexity citations. The entity being described in the post, whether a person, an institution, or a concept, gets associated with whatever language surrounds it.

For a multilateral such as UNDRR or a World Bank affiliate like CGAP, this matters in a specific way. When a programme officer posts about disaster risk metrics or financial inclusion data, the language they choose does not stay on LinkedIn. It travels into the index that AI systems draw on when answering questions about those topics. A carelessly worded post trains a model to describe the organisation in carelessly worded terms. A precise, authoritative post does the opposite.

The same logic applies to executives at industrial groups or financial institutions. If a chief sustainability officer at a firm like Holcim posts consistently about decarbonisation in construction, the entity-recognition systems that feed AI answers begin to associate that person's name with that topic. The post is doing work that a press release cannot: it is building a machine-readable record of expertise.

What actually drives ranking and recognition

Sendible Insights identifies three conditions that determine whether a LinkedIn post reaches Google's index and influences AI models. First, the post must be public. Second, it needs meaningful early engagement, saves and substantive comments in particular, because these signal to both LinkedIn's own distribution system and to crawlers that the content is worth indexing. Third, the language must be specific: named entities, precise claims, and subject-matter terminology that a language model can map to a known topic.

Generic content fails all three tests. A post that says "excited to share our latest thinking on resilience" gives a crawler nothing to work with. A post that says "UNDRR's 2025 data shows a 34% gap in early warning coverage across sub-Saharan Africa" gives it a named institution, a year, a figure, and a topic. One of these posts contributes to an organisation's AI-visible authority. The other does not.

The implication for posting cadence is also concrete. Frequency matters less than consistency of topic. A single post on disaster finance published by someone who otherwise posts about weekend hiking contributes almost nothing to entity recognition. Forty posts over six months, all anchored to the same domain, build a pattern that both Google and LLMs can read as expertise. This is the LinkedIn equivalent of the topical authority that SEO practitioners build on websites, except it accrues to a person rather than a domain.

The posts that are already doing this work, unintentionally

Most B2B organisations already have executives posting on LinkedIn. A significant fraction of those posts are public. Some subset earns enough engagement to be indexed. The question is whether the language in those posts is doing the entity-recognition work it could be doing, or whether it is being squandered on vague affirmations and conference check-ins.

A financial services firm whose managing director posts "great panel today on digital assets" is losing the indexing opportunity that "BlackRock's tokenisation pilot raises a custody question the Basel framework does not yet answer" would capture. The first post is invisible to AI systems. The second begins to associate a name with a specific, searchable claim.

The Pedro Dias example in Sendible's piece is deliberately absurd, but it is also the cleanest possible proof of mechanism. If a self-satirising claim posted on LinkedIn can populate a Google AI Overview within hours, the same infrastructure is available to any executive who posts a substantive, specific, well-timed claim in their domain. The difference is that the executive's claim would actually be worth making.

B2B authority has always been built by saying the right things in the right places. LinkedIn is now one of the places where what you say gets recorded, ranked, and repeated by systems that an increasing share of your buyers consult before they ever visit your website. Post accordingly.