Schema Markup for AI Search: JSON-LD Guide for Small Businesses

Schema markup is structured data you add to your website\'s HTML that tells search engines and AI systems exactly what your business is, what you offer, and how to find you. While human visitors read your website visually — scanning headings, images, and text — AI engines parse the underlying code. Without schema markup, AI engines must guess the meaning of your content from unstructured text. With it, they can extract precise, machine-readable facts: your business name, type, address, hours, services, prices, and reviews. This precision matters enormously in the age of AI search. When ChatGPT, Perplexity, or Google AI Overviews generate a recommendation, they need to be confident in the facts they present. Structured data provides that confidence.

• LocalBusiness (or specific subtype)
• Organization
• FAQPage
• HowTo
• Article with author and publisher
• BreadcrumbList
• AggregateRating

Why specificity matters in schema types

A website that declares itself a "Dentist" in LocalBusiness schema, with a verified address, specific service list, and aggregate review rating, gives the AI far more to work with than a website that simply mentions "dental services" somewhere in a paragraph. The numbers make the case clearly. ChatGPT reaches 2.8 billion monthly active users. Google AI Overviews appear in 47 percent of search results. Together, these platforms represent a massive audience that is actively asking AI engines for business recommendations. If your website does not speak the AI\'s language through structured data, you are invisible to this audience — regardless of how good your actual services are. But there is a critical nuance that most guides miss: schema quality matters more than schema presence. According to research from White Hat SEO, empty or generic schema markup can result in an 18 percent citation penalty compared to having no schema at all. Adding placeholder schema with incomplete or inaccurate information is actively harmful. Every schema element on your website should contain accurate, specific, and current data about your business.

For most small and local businesses, two schema types form the foundation of your structured data strategy: LocalBusiness and Organization. These are the schemas that tell AI engines the fundamental facts about your business identity. LocalBusiness schema is the most important type for any business that serves customers in a specific geographic area. This includes restaurants, dentists, plumbers, law firms, auto repair shops, salons, and any other business with a physical location or service area. LocalBusiness schema declares your business name, address, phone number, hours of operation, price range, payment methods accepted, geographic service area, and the specific type of business you are (using subtypes like Restaurant, Dentist, LegalService, or Plumber for greater specificity). Organization schema establishes your brand identity at a higher level.

Required Organization schema properties

It includes your business name, logo URL, founding date, founding location, description, social media profile links, and contact information. While LocalBusiness inherits from Organization, having both provides AI engines with complementary data. Organization schema is particularly important for businesses that operate online or across multiple locations, where the brand identity extends beyond a single physical address. The key to effective LocalBusiness and Organization schema is specificity. Do not use the generic "LocalBusiness" type when a more specific subtype exists. A dentist should use the "Dentist" type. A plumber should use "Plumber." A restaurant should use "Restaurant" and include cuisine type. These specific subtypes help AI engines match your business to precise queries. When someone asks ChatGPT for "the best Italian restaurant in downtown Calgary," the AI gives preference to businesses that have declared themselves as restaurants with Italian cuisine through schema markup, because that structured data eliminates guesswork. Include your complete address using PostalAddress schema with street address, city, state or province, postal code, and country. Include GeoCoordinates with your exact latitude and longitude. AI engines use these coordinates to verify your location and match you to location-specific queries.

If you implement only one type of schema markup beyond LocalBusiness, make it FAQPage. FAQ schema is arguably the highest-impact structured data type for AI visibility because it directly matches the question-and-answer format that AI engines use to generate responses. When a customer asks ChatGPT "how much do dental implants cost in Toronto?" the AI looks for sources that provide a direct, structured answer to that specific question. FAQPage schema wraps your questions and answers in a format that AI engines can parse instantly, making your content dramatically easier to cite. According to CXL, FAQ content is one of the most frequently cited content types in AI-generated answers. This makes intuitive sense: AI engines are essentially answering questions, so content that is already structured as questions and answers provides a ready-made building block for the AI\'s response.

How FAQPage schema feeds AI answers

FAQPage schema takes each question-answer pair on your page and marks it up so the AI can extract it as a discrete, citable unit. Implementation is straightforward. Each FAQ page on your website should have a single FAQPage schema block that contains all the question-answer pairs on that page. Each pair uses the Question type with a "name" property containing the question text and an AcceptedAnswer containing the answer text. The answers should be concise but specific — include price ranges, timelines, qualifications, and what makes your approach different. Avoid generic answers that could apply to any competitor. The most effective FAQ content comes from your real customers. Review the questions you receive through phone calls, emails, contact forms, Google Business Profile Q&A, and social media. These are the exact questions potential customers are asking AI engines about your industry. Create dedicated FAQ sections on each service page addressing the 5 to 10 most common questions about that specific service. Then mark each section up with FAQPage schema. The combination of real customer questions, specific expert answers, and proper schema markup is one of the most powerful AI visibility improvements available.

Beyond business identity and FAQ content, two additional schema types significantly boost your AI visibility for content-rich websites: Article schema and HowTo schema. These types help AI engines understand and cite your educational and instructional content. Article schema should be applied to blog posts, guides, news articles, and any long-form content on your website. It declares the article\'s headline, author, publication date, modification date, description, and publisher. Author schema within articles is particularly important for AI visibility because AI engines evaluate the authoritativeness of content partly based on who wrote it. Include the author\'s name, job title, credentials, and a link to their bio page. This creates an entity relationship between your expert authors and your business that strengthens AI\'s confidence in citing your content.

When to use HowTo schema

For businesses that provide step-by-step services or instructional content, HowTo schema is highly valuable. A cleaning company might mark up "How to remove red wine stains from carpet." A financial advisor might mark up "How to create a monthly household budget." A dentist might mark up "How to prepare for a root canal procedure." Each step in the HowTo includes a name, description, and optionally an image and time estimate. HowTo schema is particularly effective for AI citation because it structures information in discrete, sequential steps that AI engines can reference individually. When a customer asks Perplexity "how do I prepare for a dental cleaning?" the AI can pull specific steps directly from your HowTo schema and cite your website as the source. This generates both visibility and traffic. The combination of Article schema on your blog content and HowTo schema on your instructional content tells AI engines that your website is not just a business listing — it is an authoritative source of expert knowledge in your field. This perception of expertise directly influences how frequently and prominently AI engines cite your content.

For businesses that sell products or offer services with clear pricing, Product and Review schema types create the structured data that turns AI mentions into actual sales conversations. Product schema declares your product or service name, description, price, currency, availability, brand, and images. When a customer asks an AI engine "how much does carpet cleaning cost in Winnipeg?" a business with Product schema that includes specific pricing for carpet cleaning services gives the AI a precise, citable answer. Without Product schema, the AI must estimate from unstructured text or skip your business entirely. AggregateRating schema works alongside Product and LocalBusiness schema to surface your reputation data. It declares your average star rating, total review count, and the scale used. This structured reputation data gives AI engines a quick, quantified trust signal.

How review schema boosts AI confidence

A business with an AggregateRating showing 4.7 stars across 142 reviews provides stronger confidence than a business with no structured review data, even if that second business has equally good reviews — the AI simply cannot parse them as efficiently. Individual Review schema allows you to mark up specific customer reviews on your website. Each review includes the author\'s name, the rating value, the review body text, and the date. While Google\'s guidelines restrict first-party review markup in some contexts, having structured review data on your website still helps AI engines understand the sentiment and volume of customer feedback. The most effective approach is maintaining reviews on trusted third-party platforms (Google Business Profile, Yelp, industry-specific review sites) while using AggregateRating schema on your website to summarize your overall reputation. AI engines cross-reference these signals: if your website declares 4.7 stars across 142 reviews and Google Business Profile shows a matching rating, the AI\'s confidence in recommending you increases significantly. If the numbers do not match, it raises a trust flag. This is why accuracy in schema markup is non-negotiable — every data point must reflect reality.

Schema markup implementation varies by platform, but the underlying principle is the same: you are adding JSON-LD code blocks to the HTML of your web pages. JSON-LD (JavaScript Object Notation for Linked Data) is the format recommended by Google and preferred by AI engines because it lives in a script tag that does not affect your visible page content. For WordPress websites, which power approximately 40 percent of the web, you have three options. First, you can paste JSON-LD directly into your theme\'s header.php file or use the "Additional Header Scripts" section available in many themes. Second, you can use a plugin like Insert Headers and Footers, WPCode, or Schema Pro to manage your structured data. Third, SEO plugins like Yoast and RankMath include built-in schema generators with varying levels of customization.

Platform-specific implementation guides

For Squarespace websites, navigate to Settings, then Advanced, then Code Injection, and paste your JSON-LD into the Header section. This applies the schema globally. For page-specific schema, use the Page Settings panel for each page and add the code to that page\'s header injection area. For Wix websites, go to Settings, then Custom Code, and add your JSON-LD as a new code snippet set to load in the Head section. Wix also offers an SEO panel on each page where you can add structured data. For custom HTML websites, simply add a script tag with type "application/ld+json" in the head section of each page. Place your JSON-LD object inside the script tag. LunimRank\'s free Schema Generator produces properly formatted JSON-LD ready to paste. It handles the technical details that commonly cause implementation errors: proper ISO date formatting for business hours, correct address nesting with PostalAddress type, valid geographic coordinates, and proper array formatting for multiple items like service offerings and accepted payment methods. After generating your schema, the tool provides platform-specific implementation instructions.

Implementing schema markup without validating it is like filing your taxes without checking the math. Errors in schema markup not only fail to help your AI visibility — they can actively hurt it through the 18 percent citation penalty that White Hat SEO documented for empty or incorrect schema. Validation is a mandatory step, not an optional one. Three validation tools should be part of your workflow. Google\'s Rich Results Test at search.google.com/test/rich-results validates your schema against Google\'s specific requirements. It shows which rich result types your page is eligible for and flags any errors or warnings. This tool is essential because Google AI Overviews — which reach 1.5 billion monthly users — use the same structured data that Google\'s Rich Results system uses. If your schema passes this test, Google\'s AI can read it correctly.

Tools for testing your schema markup

The Schema.org Validator at validator.schema.org provides the most detailed validation against the full schema.org specification. It catches issues that Google\'s test might miss, including deprecated properties, recommended versus required field distinctions, and type hierarchy violations. Use this as your second validation layer. Google Search Console\'s Enhancements tab shows schema errors and warnings across your entire website over time. Unlike the point-in-time tests above, Search Console monitors your structured data continuously and alerts you to new issues. Check it weekly to catch problems early. LunimRank\'s free Schema Validator provides a fourth layer of validation specifically focused on AI visibility. It checks your schema against the factors that AI engines prioritize, including completeness of business information, presence of recommended properties (not just required ones), and consistency with your other online listings. Common validation errors include missing required properties like "@type" or "name," invalid URL formats in image or logo fields, opening hours in the wrong format (ISO 8601 durations versus plain text), and missing address components. The Schema Generator avoids most of these by pre-filling the correct format, but always validate after implementation because copy-paste errors and CMS formatting can introduce issues.

Understanding what not to do with schema markup is as important as knowing what to do. These are the most common mistakes we see at LunimRank, each of which can reduce your AI visibility rather than improve it. The first mistake is generic or placeholder schema. Adding LocalBusiness schema with "Your Business Name" or "123 Main Street" as placeholder text and forgetting to update it is worse than having no schema at all. AI engines detect placeholder content and apply a citation penalty. Every field in your schema must contain your actual, accurate business information. The second mistake is inconsistent data between schema and visible page content. If your schema declares business hours of 9 AM to 5 PM but your visible page content says 8 AM to 6 PM, AI engines lose trust in both sources. Schema markup should exactly match the information displayed on your page.

Avoiding generic and outdated schema

The third mistake is using the wrong schema type. A law firm using generic LocalBusiness instead of LegalService, or a dentist using MedicalBusiness instead of Dentist, misses opportunities for precise matching. Always use the most specific schema type available for your business. The fourth mistake is adding schema for services you do not actually offer. Some businesses add extensive service lists to their schema to appear for more queries. AI engines cross-reference this data against your actual page content and other online listings. Schema that does not match reality triggers trust penalties. The fifth mistake is implementing schema on only one page. Your homepage schema is important, but AI engines evaluate your entire site. Each service page should have its own schema relevant to that service. Each blog post should have Article schema. Each FAQ section should have FAQPage schema. Comprehensive schema coverage across your entire site creates a much stronger AI visibility profile than a single schema block on your homepage. The sixth mistake is never updating schema after initial implementation. If your business hours change, your phone number changes, or you add new services, your schema must be updated to match. Set a quarterly calendar reminder to review and update all schema markup on your website.

Once you have the foundational schema types in place — LocalBusiness, FAQPage, Article, and Product — several advanced strategies can further boost your AI visibility. BreadcrumbList schema tells AI engines how your website is organized. It maps the navigation path from your homepage to each page, helping AI understand the relationship between your content. For example: Home > Services > Residential Plumbing > Drain Cleaning. This hierarchy helps AI engines understand that your drain cleaning page is a specialized subset of your plumbing services, which adds context when the AI generates recommendations. SameAs property links your business entity across platforms. In your Organization or LocalBusiness schema, the SameAs property should list URLs for your Google Business Profile, Facebook page, LinkedIn page, Twitter profile, Instagram profile, and any industry directories where you appear.

Building entity relationships with sameAs

This creates a web of entity connections that helps AI engines confidently identify your business across sources. It is a direct boost to entity recognition, which is a prerequisite for being recommended. ServiceArea schema specifies the geographic regions you serve. For businesses that travel to customers — plumbers, electricians, cleaning services, landscapers — ServiceArea schema tells AI engines exactly which neighborhoods, cities, or regions you cover. When someone asks ChatGPT for "a plumber that serves the east end of Winnipeg," a business with ServiceArea schema listing that specific area has a matching advantage. Event schema is valuable for businesses that host events, classes, workshops, or appointments. Restaurants can mark up tasting events. Yoga studios can mark up class schedules. Professional services can mark up webinars. Event schema gives AI engines time-specific information that makes your business relevant to "what is happening this weekend" or "upcoming yoga classes near me" queries. OfferCatalog schema groups multiple service offerings together with pricing, allowing AI engines to understand your complete service portfolio at a glance. This is particularly effective for businesses with tiered pricing or bundled service packages.

Schema markup is one of the highest-ROI investments in AI visibility. The work is front-loaded — once implemented and validated, schema requires only periodic updates to maintain. Here is your step-by-step implementation plan. Step 1: Use LunimRank\'s free Schema Generator to create LocalBusiness or Organization JSON-LD for your homepage. Fill in every field with accurate, specific information. Select the most specific business subtype available. Download or copy the generated code. Step 2: Implement the homepage schema on your website using the platform-specific instructions provided by the generator. Validate using Google\'s Rich Results Test. Fix any errors before proceeding. Step 3: Create FAQPage schema for each service page. Use real customer questions and specific, expert answers.

Your schema implementation timeline

Implement and validate each page individually. Step 4: Add Article schema to your blog posts and content pages. Include author information with credentials and job titles. Step 5: If applicable, add Product schema with pricing, HowTo schema for instructional content, and AggregateRating schema summarizing your reviews. Step 6: Add SameAs properties linking to all your social and directory profiles. Add ServiceArea schema if you serve a specific geographic region. Step 7: Validate your entire website\'s schema through Google Search Console\'s Enhancements tab and LunimRank\'s Schema Validator. Fix any warnings, not just errors. Step 8: Set a quarterly calendar reminder to review all schema markup. Update any information that has changed — hours, phone numbers, services, pricing, review counts. After implementing comprehensive schema markup, run a free LunimRank scan to see how your AI Readiness Score has changed. Most businesses see a 15 to 25 point improvement in their score from schema implementation alone. For ongoing monitoring that tracks your schema\'s impact on AI visibility over time, LunimRank\'s paid plans starting at 39 dollars per month include weekly automated scans and competitor benchmarking.