A company’s website says its team sets up equipment on site in Dubai. Its markup says the service covers the whole UAE. A plugin also adds a field that claims the page is eligible for AI answers. The page has two independent problems. The coverage statement is broader than the company approved, and the eligibility field has no basis in the vocabulary being used. Fixing the syntax alone leaves the coverage error in place.
If you are deciding whether to add structured data to a service page, first identify the service and the facts the company can maintain. Then name the system or feature you expect to use the markup and look for its documentation. The existing AI Visibility service-page guide explains how to write copy that answers customer questions. This walkthrough addresses the implementation review after that copy exists. AI Visibility’s service-page writing guide
Start with the service the visitor sees
Open the page and name what the customer can enquire about, reserve or buy. The answer might be on-site installation, an annual maintenance visit or a remote training session. Name the provider separately. One organization can deliver several services, and a service can be available through more than one channel. Preserve those distinctions when mapping the page to structured data.
A service page can also mention a physical office or warehouse. That address tells a reader where the business operates from, while service coverage tells them where the work can be provided. An office in Dubai does not establish that a technician travels to every address in Dubai. A company with offices in two emirates may still offer a particular service only in one of them. Confirm coverage with the service owner and make it clear in the visible text before carrying it into markup.
| Field | What to confirm in the page and source record | When to pause |
|---|---|---|
| Service name and type | What the provider actually offers | The page describes the whole company or several unrelated services |
| Provider | The organization or person responsible for delivery | The brand and the contracting provider are unclear |
| Service area | The approved places where this service is offered | The draft expands coverage beyond the visible copy |
| Access channel | The stated way to reach or request the service | A contact page is being treated as confirmed booking |
| Offer details | Published terms that apply to the particular service | A quote or conditional price is presented as a fixed amount |
For a bilingual site, capture the approved version of the facts in both languages. Translation can introduce a broader service label or replace a city with the country name. Review the two public pages with the person who owns the offering. If the same service is represented in both versions, preserve its identity and confirm that each translation communicates the same underlying scope.
Use the vocabulary to describe real facts
Schema.org defines Service and includes properties such as serviceType, provider and areaServed. It also defines availableChannel for a way to access a service. These terms can express relationships when they describe the page correctly. For instance, an on-site setup service can identify its provider and the city where staff deliver it. A separately described location can carry location-specific facts. Schema.org Service
Add optional details when the page and operations record support them. A page that invites readers to request a quote can say how the quote works and which details affect it. If there is no approved public price, skip a numeric price rather than inserting zero as a placeholder. For an offer limited by location or appointment terms, preserve those conditions in the copy and confirm the semantics of any offer properties before publishing. Good markup gives a reader or downstream system fewer opportunities to assign a fact to the wrong service.
When the company has a shared contact form, check where it routes enquiries. A form can be a way to reach the team even if it never checks calendar availability. An appointment calendar can show selectable times without confirming a booking until a later step. Identify the actual function in the copy and review any machine-readable channel or action against what the site does. Leave undocumented capabilities out of the markup.
Check which system is expected to use it
Use three separate questions in the release review. Does the vocabulary express the intended service facts? Does the named search feature document support for that type of markup? Does the published page actually appear in the feature? The first is a vocabulary and content question. The second comes from the platform’s documentation and requirements. The third needs a real observation from that platform. A yes to one question does not settle the others.
As checked on 6 November 2026, Google’s Search structured-data gallery did not document a standalone rich result for generic Service markup. A service page may contain other entities, such as a breadcrumb trail or a physical branch, and those have their own documentation and requirements. Review each one for what it actually represents. Relabelling the service as another entity just to get a validator result can introduce a factual error in an otherwise accurate page. Google Search structured-data gallery · Google structured-data policies
Google also says its generative AI features in Search do not require special Schema.org markup. Its guidance points website owners to ordinary Search requirements and eligibility for a snippet when assessing whether a page could appear as a supporting link. Those conditions concern Google Search. If a plugin or supplier promises a different platform will use a field, ask for the name of that platform, its current documentation and the outcome the supplier can actually verify. Google AI features guidance
An official DESCO Print & Studio Reel illustrates the difference between a source that describes a service and one that confirms its scope. The caption promotes finishing services across UAE branches. It provides a published description from the company. An editor adding service markup for one branch would still need that branch’s own approved service information before asserting availability there. Record which entity, offer and period each source covers. Official DESCO Instagram Reel
Test a service claim before and after correction
On 6 November 2026, a local fixture check compared two fictional Service nodes with a Schema.org vocabulary capture and approved page facts. The approved facts described on-site equipment setup from Example Workshop in Dubai. The first fixture expanded areaServed to United Arab Emirates and added the invented property aiAnswerEligible. It was still valid JSON.
The local checker reported the invented field as unknown in the captured vocabulary. The separate fact comparison flagged the broader coverage statement. areaServed accepts a text value in this context, so the text naming the country passed the vocabulary check. Checking the approved page facts caught that error. The corrected fixture narrowed the area to Dubai and removed the invented field. Both sets of local checks then returned zero findings. Schema.org vocabulary distribution
{
"@context": "https://schema.org",
"@type": "Service",
"name": "On-site equipment setup",
"serviceType": "On-site equipment setup",
"provider": {
"@type": "Organization",
"name": "Example Workshop"
},
"areaServed": "Dubai"
}
This fixture shows the fields involved in the check. It represents a fictional service and does not set a minimum field list for Google. The checker used selected Schema.org term, domain and range declarations and compared values with a short reference record. It did not run a complete JSON-LD processor or submit the fixture to a search platform. On a live site, compare the markup with the rendered page and approved records as well. A parity check can otherwise approve two copies of the same stale statement.
Choose the right validation and release checks
Google documents two different tools. Schema Markup Validator checks Schema.org markup in general, while Google’s Rich Results Test is intended to show which Google rich results the submitted markup could support. Record which tool you ran and what it reported. When a type falls outside the rich-result feature you are testing, use the vocabulary validator and the page comparison to assess implementation quality. Track any supported features on the same page separately. Google’s structured-data testing guide
Give the developer a Service markup record with the page, service name, provider, area and access channel. Put the visible wording next to each structured value, name its approving owner and record when it was checked. Add separate fields for vocabulary results, documented platform support and any observed appearance. This lets a reviewer see immediately whether an implementation check has been presented as a claim about search visibility.
Inspect the deployed output after publishing. Check the template and plugins for a second node that uses a broader location or an old provider name. A business with several service pages can accidentally emit the company’s whole coverage area on every page. Test two services with deliberately different coverage. If a shared component generates the same value for both, trace it to the CMS field or default that supplied it and update that source. Check an unaffected page too, especially if the correction changes a shared template.
Ask the operations owner to tell the web team when coverage changes, a service pauses or a provider takes over. Give each change an effective date. Review the visible copy, enquiry route and emitted markup after the update. If the provider still answers enquiries but no longer visits a district, update the scope while preserving the contact path customers can use. A second editor should be able to repeat the check using the saved page, expected statement and affected structured field.
If the team is choosing between rewriting unclear copy and adding more fields, start with the customer’s missing answer. A reliable description of what is included, where the team works and how to get a quote gives the editor a source for the structured values. Add markup when its purpose is clear and someone can keep it current. An optional property that nobody can verify is a poor release requirement.
For a rollout across many pages, sample the cases that are easiest to confuse. Pick an on-site service with a narrow delivery area, a remote service with different terms and a branch that offers only part of the catalogue. Record the responsible provider and available channel for each. Test those pages again after a template change and after the first real coverage update.
Treat a proposed field from a supplier as an implementation change you can review. Ask for the exact property name, the type it belongs on and a reference for the feature the supplier expects it to affect. If the supplier is describing a proprietary format, keep it separate from Schema.org and verify that the intended system consumes it. If you cannot identify that system or a maintained use, keep the release focused on the verified service facts.
AI Visibility is published by Lunasol, which offers GEO and AEO services. The technical example used fictional Service nodes and a local vocabulary comparison. About AI Visibility



