As entrepreneurs, we love an excellent funnel. It supplies readability on how our methods are working. We now have conversion charges and may observe the client journey from discovery by way of conversion. However in at the moment’s AI-first world, our funnel has gone darkish.
We will’t but absolutely measure visibility in AI experiences like ChatGPT or Perplexity. Whereas rising instruments provide partial insights, their information isn’t complete or constantly dependable. Conventional metrics like impressions and clicks nonetheless don’t inform the entire story in these areas, leaving entrepreneurs dealing with a brand new sort of measurement hole.
To assist deliver readability, let’s have a look at what we all know and don’t learn about measuring the worth of structured information (also called schema markup). By understanding either side, we will give attention to what’s measurable and controllable at the moment, and the place the alternatives lie as AI adjustments how clients uncover and interact with our manufacturers.
Why Most ‘AI Visibility’ Information Isn’t Actual
AI has created a starvation for metrics. Entrepreneurs, determined to quantify what’s occurring on the high of the funnel, are turning to a wave of recent instruments. Many of those platforms are creating novel measurements, comparable to “model authority on AI platforms,” that aren’t grounded in consultant information.
For instance, some instruments are attempting to measure “AI prompts” by treating quick key phrase phrases as in the event that they have been equal to shopper queries in ChatGPT or Perplexity. However this method is deceptive. Shoppers are writing longer, context-rich prompts that go far past what keyword-based metrics recommend. These prompts are nuanced, conversational, and extremely customized – nothing like conventional long-tail queries.
These artificial metrics provide false consolation. They distract from what’s truly measurable and controllable. The actual fact is, ChatGPT, Perplexity, and even Google’s AI Overviews aren’t offering us with clear and complete visibility information.
So, what can we measure that really impacts visibility? Structured information.
What Is AI Search Visibility?
Earlier than diving into metrics, it’s value defining “AI search visibility.” In conventional search engine optimisation, visibility meant showing on web page one in every of search outcomes or incomes clicks. In an AI-driven world, visibility means being understood, trusted, and referenced by each engines like google and AI methods. Structured information performs a task on this evolution. It helps outline, join, and make clear your model’s digital entities in order that engines like google and AI methods can perceive them.
The Knowns: What We Can Measure With Confidence For Structured Information
Let’s discuss what is thought and measurable at the moment with regard to structured information.
Elevated Click on-By means of Charges From Wealthy Outcomes
From information in our quarterly enterprise assessment, we see, by implementing structured information on a web page, the content material qualifies for a wealthy outcome, and enterprise manufacturers constantly see a rise in click-through charges. Google at present helps greater than 30 varieties of wealthy outcomes, which proceed to seem in natural search.
For instance, from our inside information, in Q3 2025, one enterprise model within the residence home equipment trade noticed click-through charges on product pages enhance by 300% when a wealthy outcome was awarded. Wealthy outcomes proceed to supply each visibility and conversion good points from natural search.
Elevated Non-Branded Clicks From Sturdy Entity Linking
It’s essential to tell apart between primary schema markup and sturdy schema markup with entity linking that leads to a data graph. Schema markup describes what’s on a web page. Entity linking connects these issues to different well-defined entities throughout your web site and the net, creating relationships that outline which means and context.
An entity is a singular and distinguishable factor or idea, comparable to an individual, product, or service. Entity linking defines how these entities relate to at least one one other, both by way of exterior authoritative sources like Wikidata and Google’s data graph or your individual inside content material data graph.
For instance, think about a web page a couple of doctor. The schema markup would describe the doctor. Sturdy, semantic markup would additionally hook up with Wikidata and Google’s data graph to outline their specialty, whereas linking to the hospital and medical providers they supply.
Picture from writer, November 2025AIO Visibility
Conventional search engine optimisation metrics can’t but measure AI experiences instantly, however some platforms can establish some cases when a model is talked about in an AI Overview (AIO) outcome.
Analysis from a BrightEdge report discovered that adopting entity-based search engine optimisation practices helps stronger AI visibility. The report famous:
“AI prioritizes content material from recognized, trusted entities. Cease optimizing for fragmented key phrases and begin constructing complete matter authority. Our information reveals that authoritative content material is thrice extra prone to be cited in AI responses than narrowly targeted pages.”
The Unknowns: What We Can’t But Measure
Whereas we will measure the influence of entities in schema markup by way of current search engine optimisation metrics, we don’t but have direct visibility into how these parts affect giant language mannequin (LLM) efficiency.
How LLMs Are Utilizing Schema Markup
Visibility begins with understanding – and understanding begins with structured information.
Proof for that is rising. In Microsoft’s Oct. 8, 2025 weblog put up, “Optimizing Your Content material for Inclusion in AI Search Solutions (Microsoft Promoting,” Krishna Madhaven, Principal Product Supervisor for Microsoft Bing, wrote:
“For entrepreneurs, the problem is ensuring their content material is straightforward to know and structured in a approach that AI methods can use.”
He added:
“Schema is a sort of code that helps engines like google and AI methods perceive your content material.”
Equally, Google’s article, “High methods to make sure your content material performs properly in Google’s AI experiences on Search,” reinforces that “structured information is beneficial for sharing details about your content material in a machine-readable approach.”
Why are Google and Microsoft each emphasizing structured information? One motive could also be value and effectivity. Structured information helps construct data graphs, which function the muse for extra correct, explainable, and reliable AI. Analysis has proven that data graphs can cut back hallucinations and enhance efficiency in LLMs:
Whereas schema markup itself isn’t usually ingested instantly to coach LLMs, the retrieval section in retrieval-augmented technology (RAG) methods performs an important position in how LLMs reply to queries. In latest work, Microsoft’s GraphRAG system generates a data graph (through entity and relation extraction) from textual information and leverages that graph in its retrieval pipeline. Of their experiments, GraphRAG usually outperforms a baseline RAG method, particularly for duties requiring multi-hop reasoning or grounding throughout disparate entities.
This helps clarify why firms like Google and Microsoft are encouraging enterprise manufacturers to put money into structured information – it’s the connective tissue that helps AI methods retrieve correct, contextual info.
Past Web page-Degree search engine optimisation: Constructing Information Graphs
There’s an essential distinction between optimizing a single web page for search engine optimisation and constructing a data graph that connects your complete enterprise’s content material. In a latest interview with Robby Stein, VP of Product at Google, it was famous that AI queries can contain dozens of subqueries behind the scenes (often called question fan-out). This means a degree of complexity that calls for a extra holistic method.
To achieve this setting, manufacturers should transfer past optimizing pages and as a substitute construct data graphs, or quite, a knowledge layer that represents the complete context of their enterprise.
The Semantic Internet Imaginative and prescient, Realized
What’s actually thrilling is that the imaginative and prescient for the semantic net is right here. As Tim Berners-Lee, Ora Lassila, and James Hendler wrote in “The Semantic Internet” (Scientific American, 2001):
“The Semantic Internet will allow machines to understand semantic paperwork and information, and allow software program brokers roaming from web page to web page to execute refined duties for customers.”
We’re seeing this unfold at the moment, with transactions and queries occurring instantly inside AI methods like ChatGPT. Microsoft is already making ready for the subsequent stage, usually known as the “agentic net.” In November 2024, RV Guha – creator of Schema.org and now at Microsoft – introduced an open venture known as NLWeb. The aim of NLWeb is to be “the quickest and best option to successfully flip your web site into an AI app, permitting customers to question the contents of the location by instantly utilizing pure language, similar to with an AI assistant or Copilot.”
In a latest dialog I had with Guha, he shared that NLWeb’s imaginative and prescient is to be the endpoint for brokers to work together with web sites. NLWeb will use structured information to do that:
“NLWeb leverages semi-structured codecs like Schema.org…to create pure language interfaces usable by each people and AI brokers.”
Turning The Darkish Funnel Into An Clever One
Simply as we lack actual metrics for measuring model efficiency in ChatGPT and Perplexity, we additionally don’t but have full metrics for schema markup’s position in AI visibility. However we do have clear, constant indicators from Google and Microsoft that their AI experiences do, partially, use structured information to know content material.
The way forward for advertising belongs to manufacturers which might be each understood and trusted by machines. Structured information is one issue in the direction of making that occur.
Extra Sources:
Featured Picture: Roman Samborskyi/Shutterstock
