Semantic Triplet Checker

The clearest sentences express one idea, using a subject, a verb and an object. For example: “AI visibility measures how often AI assistants mention a brand.” This structure is called a semantic triplet, and it makes the meaning of a sentence unusually explicit. That matters for visibility in AI search because AI systems are constantly trying to work out what things are, how they relate to one another and which facts are safe to use in an answer. Paste in a sentence to see how clear yours is.

What’s a semantic triplet?

A semantic triplet is the simplest way to state a fact. It has three parts:

Entity → relationship → entity or value

For example: “AI visibility → measures → how often AI assistants mention a brand”, or “The Guardian → was first published in → 1821”. In a sentence, the three parts are usually the subject, the verb and the object. Knowledge graphs, like the one behind Google’s knowledge panels, store facts in this shape.

Why writers should care

A clear triplet means nobody has to work out who did what, because the sentence says it. It helps AI systems in three ways:

• It names the thing. “Clarivate” is an entity an AI system can recognise and link to what it already knows. “We” or “our platform” means nothing once the sentence is lifted out of the page.
• It states the relationship. Verbs like “owns”, “measures” and “was founded in” say exactly how two things connect. “Helps with” or “is committed to” leaves the system to guess.
• It gives a fact that can be checked. A name, a number or a date can be compared with other sources. A fact that other sources confirm may be safer for an AI system to repeat than a vague claim it can’t check.

AI assistants often quote or paraphrase a single sentence or passage rather than a whole page. A sentence that carries its own subject, relationship and object still makes sense on its own, so it’s easier to use in an answer.

No search engine or AI company has said it pulls triplets out of web pages or ranks sentences by them. Treat the idea as a way of writing clearly, not as a ranking factor. The advice holds up anyway: a sentence that names its subject and says something specific is easier to understand, easier to match to a question and easier to quote.

What this tool checks

For each sentence, it picks out the subject, the verb and the object, then checks:

• The subject is named. “It” or “We” is fine once the paragraph has named the subject. If a paragraph never names it, an AI tool that quotes that paragraph on its own won’t know what it’s about.
• The verb is specific. Verbs like “helps”, “offers” and “provides” are flagged when what follows is vague too.
• The object is concrete. Phrases such as “transformative intelligence”, “workflow solutions” or “a range of things” are flagged. A number or a name usually fixes them.
• One idea per sentence. A second claim tacked on with “and”, “which” or “but” is flagged, as are sentences over 25 words.
• Passive voice, hedging and filler. Hedges aren’t always wrong. Keep one if the claim genuinely isn’t certain.

What it can’t tell you

It can’t tell whether a sentence is true, whether it says anything new, or whether it answers a question people actually ask. Those probably matter more for being cited by AI tools than sentence structure does. It’s also a set of rules, not a language model, so it will sometimes get things wrong. If it does, use the “Not right?” links to tell me.

Need help with it?

I run content strategy and AI visibility work for publishers and brands, and training for editorial teams. Book a call or get in touch.