Schema can make an existing fact explicit. It cannot turn an unsupported claim into evidence or force an AI engine to cite a page.
There are no eight universally required “AI schema” types. The useful selection depends on what the page actually represents.
Eight types to consider
| Type | Appropriate use | Frequent mistake |
|---|---|---|
| Organization | Describe the actual publishing business | Conflicting names and identifiers |
| LocalBusiness or accurate subtype | A real relevant local business, such as LegalService or AccountingService | Invented locations or credentials |
| Person | A real author, practitioner or reviewer | Creating fictional experts |
| Service | The actual service offered and its provider | Describing services never delivered |
| SoftwareApplication | A genuine software product | Applying product markup to an agency article |
| Article / BlogPosting | The editorial page itself | Fabricated authors or update dates |
| BreadcrumbList | The real navigation hierarchy | Marking up a path users cannot follow |
| WebPage | Identify the page and relationships in the graph | Treating a valid graph as proof of a rich result |
Subtypes are choices within these categories; the table is not a ranking of citation power.
Connect the entities
Use stable @id identifiers so the organization referenced as publisher is the same organization described elsewhere. Keep relevant sameAs URLs limited to genuine official identity profiles.
Example: an agency blog post
The article can identify its actual author and Buzzlect as publisher. An article about accounting should not declare Buzzlect to be an AccountingService.
Validate in two different ways
Use Schema.org Validator to inspect vocabulary and relationships. Use Google’s Rich Results Test to check supported Google search enhancements. Passing either test does not prove a factual claim or guarantee display.
Compare the graph with the visible page. Check names, dates, prices, qualifications and URLs. Remove markup that describes hidden or nonexistent content.
Frequently Asked Questions
No. Useful FAQ copy can help readers without a promise of a Google enhancement or AI citation. Confirm current platform documentation before implementing feature-specific markup.
No. A Schema.org type and a supported Google search feature are different things.
Next step
Want a second pair of eyes on your structured data and entity setup?
Request a scoped technical AI visibility review →
Related guides
- Entity SEO: How Search Engines Understand Brands, People, Products and Topics
- Entity Optimization for AI Search: A Packageable Service for Agencies
- Entity SEO for B2B IT Services: A Practical Implementation Guide
- Google’s AI Content Guidelines: What B2B SEO Teams Must Know
- How ChatGPT, Perplexity and Google AI Choose Which Sources to Cite (and How to Explain It to Clients)
Sources and verification notes
The steps, examples and templates in this article are practical recommendations, not platform rules, survey data or measured client results. Hypothetical figures and examples are labelled as such. The sources below support the platform and technical facts only.
- Google: AI features and your website
- Google: Introduction to structured data markup
- Schema.org: LegalService
- Schema.org: AccountingService
- Schema.org: Service
- Schema.org: SoftwareApplication
Want to see how your brand ranks in ChatGPT right now?
Run a free AI Visibility Audit to see where your business appears across ChatGPT, Gemini and Perplexity—and where competitors are being recommended instead.
Get Your Free AI Visibility Audit