All insights
Technical21 Nov 20265 min read

The technical layer: schema, entities and citations

Markup will not get you cited on its own, but missing markup makes it harder for a model to know which entity a page belongs to. Here is the layer that is worth doing.

Structured data is often oversold as an AI visibility shortcut. It is not a shortcut. It is a disambiguation tool: it tells a machine which real-world entity a page belongs to, what the page is about, and how the pieces of your site relate.

The three types worth shipping first

Organisation. One canonical description of the business, its name, URL and sameAs links to the profiles that also describe it. This is the anchor every other claim hangs from. If the description on your site, your LinkedIn page and your directory listings disagree, this markup will not fix it, but it will stop making it worse.

Service or Product. What you offer, described in the same words used on the page. This is what lets an engine answer "who provides X" with your name attached to a specific offer rather than a vague category.

Article or FAQPage. Where you make claims, mark them up. FAQ markup is particularly useful because the question-and-answer shape mirrors how answers are assembled.

Entities are a consistency problem

Most entity problems are not technical. They are editorial: three different descriptions of the company across five places, a founding date that differs between profiles, a product renamed on the site but not in the directory.

Before touching schema, write one reference description of the business, 40 to 60 words, and use it everywhere a description is requested. Then make the same decision for each product line. This is unglamorous and it is the highest-leverage technical work available.

Citations are the other half

A model citing you needs a reason to trust the claim. On-page attribution helps: name the source, date the data, describe the method. Off-page, being referenced by sources the model already trusts matters more than any markup.

That is not a link-building programme in the old sense. It is the ordinary work of being quoted: original data, expert commentary, accurate directory records, and relationships with the publications in your field.

What to ignore

Do not chase a special file, a magic tag or a rumoured preference. There is no reliable way to buy your way into an answer, and most of the tactics sold as AI SEO are ordinary technical hygiene with a new label. Ship the basics properly, keep the entity consistent, and publish claims worth citing.