Schema markup for brand values: how to position your brand in AI assistants

TL;DR: Schema markup gives AI assistants structured, machine-readable signals about what your brand stands for. Using Organization, Brand, and ItemList schema with values-specific properties like knowsAbout and ethicsPolicy, you can shape how ChatGPT, Gemini, and Perplexity describe your company. Missing schema won't kill your citations. But it measurably lowers the odds an AI gets your positioning right.

Why do AI assistants care about schema markup at all?

AI assistants don't read your website the way a person does. They pull snapshots of your content, usually from training data, live web retrieval (Perplexity, Bing-backed Copilot, Google's AI Overviews), or both. So what decides which facts they grab and repeat? Structured data is one of the strongest signals you can send.

Schema.org markup is the shared vocabulary supported by Google, Microsoft, Yahoo, and Yandex since 2011 [1]. It tells a parser exactly what a piece of content means. A sentence that says "we care about sustainability" is mush to a machine. A JSON-LD node with @type: Organization and a knowsAbout property listing "sustainable manufacturing" is not.

Google's guidance on how AI Overviews pick sources says the system favors pages where "content is accurate, well-structured, and clearly attributed" [2]. Structured data is the most literal form of well-structured content there is.

The mechanism is the part most guides skip. When a retrieval-augmented generation system like Perplexity or Google AI Overviews fetches a live page, it processes the full HTML. Schema in the head or body gets parsed before the language model ever touches the prose. Your structured data shapes the context window the model reasons from. For brand values, that's a straight line from your markup to the model's answer.

What schema types actually affect brand positioning in AI results?

Most SEO guides stop at Organization and LocalBusiness schema. For values positioning, you need a wider set of types and, more to the point, specific properties inside them.

Here are the types that carry real weight:

Organization (schema.org/Organization): The base type. The properties that hold brand identity are knowsAbout, description, slogan, ethicsPolicy, and diversityPolicy. Almost nobody uses them, which means the signal-to-noise ratio is excellent right now.

Brand (schema.org/Brand): A sub-type of Intangible. Brand schema lets you define name, slogan, and logo as an entity separate from the Organization [10]. Pair it with the brand property on your Organization node and you build a named entity graph that AI systems can anchor to.

ItemList (schema.org/ItemList): The right tool for encoding a list of values, commitments, or principles. An AI retrieving a page with an ItemList of five named items is far more likely to list those items back than one that buries them in a paragraph.

FAQPage (schema.org/FAQPage): Questions like "What does [Brand] stand for?" or "Is [Brand] sustainable?" are exactly the conversational queries that land in AI assistants. An FAQPage block with those exact questions and tight answers pre-writes the AI's response [11].

ClaimReview (schema.org/ClaimReview): Built for fact-checkers, but brands fighting misinformation about supply chain or environmental practices can use it defensively. It signals credibility to Google's systems and to the AI consumers of Google data.

The table below maps schema types to positioning outcomes, based on how each property gets handled by structured data parsers.

| Schema Type | Key Properties | Primary AI Positioning Benefit | |---|---|---| | Organization | knowsAbout, ethicsPolicy, description | Core brand identity and values retrieval | | Brand | slogan, name, logo | Named entity disambiguation | | ItemList | itemListElement, name | Enumerating values or commitments | | FAQPage | mainEntity, acceptedAnswer | Direct Q&A capture in AI responses | | ClaimReview | claimReviewed, rating | Credibility and misinformation defense | | AboutPage | about, description | Deep context for brand background queries |

How do you write schema markup that captures brand values specifically?

The gap between generic Organization schema and values-tuned schema comes down to which properties you include and how specific you are.

Here's the difference in practice. Most brands publish something like this:

{
  "@context": "https://schema.org",
  "@type": "Organization",
  "name": "Acme Co",
  "url": "https://acme.com"
}

That tells a machine almost nothing about positioning. A values-tuned version looks more like this:

{
  "@context": "https://schema.org",
  "@type": "Organization",
  "name": "Acme Co",
  "url": "https://acme.com",
  "description": "Acme Co designs industrial tools built to last 20 years, sourced from North American suppliers with verified living-wage certification.",
  "slogan": "Built once. Built right.",
  "ethicsPolicy": "https://acme.com/ethics",
  "knowsAbout": [
    "sustainable manufacturing",
    "fair labor practices",
    "industrial tool design"
  ],
  "hasCredential": {
    "@type": "EducationalOccupationalCredential",
    "name": "B Corp Certified"
  }
}

The description property does most of the work. Google parses it and uses it for entity summaries because it's the most semantically dense field on the node [3]. Keep it under 160 characters so it extracts cleanly, and lead with your most differentiating claim.

The knowsAbout property takes an array of strings or Thing entities. Strings are fine. Be narrow, not broad. "Environmental sustainability" is weak. "Zero-waste textile manufacturing" is strong. AI systems match on specificity when the query is specific.

The ethicsPolicy property takes a URL. That page has to exist, has to be crawlable, and ideally carries its own structured data. A dead link inside a schema node is worse than leaving the property out, because some systems read a broken URL as a credibility hit.

For values you want listed explicitly, pair your Organization schema with an ItemList on the same page or your About page:

{
  "@context": "https://schema.org",
  "@type": "ItemList",
  "name": "Our Core Values",
  "itemListElement": [
    {"@type": "ListItem", "position": 1, "name": "Transparency", "description": "We publish our full supply chain annually."},
    {"@type": "ListItem", "position": 2, "name": "Durability", "description": "Every product carries a 20-year repair guarantee."}
  ]
}

This is the format AI answer engines quote directly. Each ListItem with a name and description is an extractable fact.

Schema properties present on Fortune 500 Organization schema instances

| | | |---|---| | description (any content) | 60% | | sameAs | 38% | | logo | 72% | | knowsAbout | 5% | | ethicsPolicy | 2% | | hasCredential | 3% | | diversityPolicy | 1% |

Source: Practitioner crawl analyses and Semrush structured data reports, 2023

Does schema markup actually change what ChatGPT or Claude says about a brand?

For a fixed training snapshot, no, not directly. ChatGPT's base model learned from a frozen dataset. Schema you publish today won't touch that snapshot until some future training run. That's the honest answer most schema guides dodge.

Three reasons schema still earns its keep even for models with fixed training data.

First, live retrieval is everywhere now. ChatGPT with browsing, Claude with web access, Perplexity, and Google's AI Overviews all fetch live pages. For those systems, schema on your live pages affects outputs today [2]. Perplexity discloses that it uses real-time web crawling [4], and it responds to structured data more like a search engine than a pure LLM does.

Second, Google's Knowledge Graph feeds entity data into Gemini, and schema shapes that graph. Publish consistent Organization schema with a sameAs property pointing to your Wikidata or Wikipedia entity, and you help Google's entity resolution identify you correctly [5]. Gemini pulls from this graph for factual brand claims.

Third, schema changes what other sites say about you. Press releases, industry directories, and aggregators often pick up your structured data when they republish your content. If your schema carries your values claims, those claims ride along. Training data isn't only your site. It's everything written about you.

A 2024 Authoritas study of more than 2,000 AI Overview citations found pages with complete structured data were 3.1x more likely to appear in an AI Overview than topically equivalent pages without it [6]. That's not a values-specific study, but the direction is unmistakable.

For generative engine optimization, schema is one of the highest-leverage technical moves you can make. It's crawlable and machine-readable, and it needs zero LLM interpretation to work.

What is the sameAs property and why does it matter for AI brand recognition?

The sameAs property in Organization schema is a URL, or array of URLs, pointing to other authoritative descriptions of the same entity. Wikidata, Wikipedia, Crunchbase, LinkedIn, and major industry databases all count as valid targets.

This matters a lot for AI assistants, because named entity recognition is core to how they work. When an AI reads "Patagonia", it has to resolve that to a specific entity and not mix it up with the Patagonia region of Argentina. The sameAs chain is how AI systems and knowledge graph builders confirm they've got the right entity.

No Wikidata entry or Wikipedia article? The property gets weaker, but it still helps to point at authoritative third-party profiles. A Crunchbase page, an official government business registry entry, or a major industry directory listing all give the machine something to anchor on.

The markup looks like this:

"sameAs": [
  "https://www.wikidata.org/wiki/Q12345678",
  "https://en.wikipedia.org/wiki/Acme_Co",
  "https://www.linkedin.com/company/acme-co"
]

Filling this out correctly is probably the single highest-return action in this whole playbook for brands that want Gemini to describe them accurately. Gemini's factual grounding leans hard on Google's Knowledge Graph, and that graph uses sameAs chains to merge entity records [5], [12].

Entity disambiguation sits upstream of everything else in AI search. An AI that can't confidently tell which company you are will hedge every claim it makes. That hedging shows up as vague or wrong brand descriptions.

How should you structure your About page to maximize AI values extraction?

Your About page is the most important page for AI brand positioning. It's the page retrieval systems most reliably tie to "what does this brand stand for", and it's where most brands leave the most structured data on the table.

A strong About page for AI retrieval has four layers.

The schema layer: Organization schema with every values-relevant property filled in. An ItemList for your core values. An AboutPage type on the page itself, with a description and a mainContentOfPage property pointing to the primary content block. If you want leaders tied to the brand's values, add Person schema for each with the worksFor property linking back to your Organization entity.

The prose layer: This is what the LLM reads if it ignores or can't parse your schema. Write values in plain declarative sentences. "We manufacture in Portugal using 100% renewable energy" is extractable. "We're passionate about creating a better tomorrow" is not. The first 100 words of each section carry the most weight, because retrieval systems often truncate long documents.

The evidence layer: Link to specific proof: your sustainability report (with its own structured data), your ethics policy, your certification pages. AI systems rank pages with citations and outbound evidence above bare claims. That mirrors academic source scoring, and the borrowing is deliberate.

The FAQ layer: Drop an FAQPage schema block at the bottom of your About page with the exact questions your buyers ask AI assistants about brands like yours. "Is [Brand] certified B Corp?" "Where does [Brand] manufacture?" "What are [Brand]'s environmental commitments?" Write tight, factual answers. These become the AI's default response template for those queries.

One mistake almost every brand makes: they hide values content behind a redirect or in a section rendered by JavaScript with no server-side rendering. Googlebot and most AI retrieval crawlers run limited JavaScript. Values content that lives only in client-rendered markup is invisible to schema parsers. Render your critical values content server-side.

AI SEO tools can help you audit whether your structured data is actually parsed and whether your values claims are turning up in AI-generated brand descriptions.

What mistakes break brand schema and hurt AI positioning?

Schema mistakes come in two flavors: syntax errors that stop parsers cold, and semantic errors that make AI systems distrust or ignore what they read.

Syntax errors are the easy fix. Run every page through Google's Rich Results Test [7] before you ship. The usual suspects: trailing commas in JSON-LD arrays, wrong property nesting, using properties that don't exist on a given @type, and unclosed quotation marks. One syntax error can void the entire JSON-LD block.

Semantic errors are subtler and do more damage to AI positioning.

Schema that contradicts the page. If your ethicsPolicy points to a URL that returns a 404, or your description says "certified B Corp" with nothing on the page to back it, AI systems that cross-reference claims will catch the mismatch. Google's quality rater guidelines cover schema-content consistency as a quality factor [8].

Vague or empty description fields. The description is too often left as the company tagline, which says nothing. "Innovating for a better future" tells an AI zero. If your description doesn't hold at least one specific, verifiable claim, it's noise.

Duplicate schema blocks. Some CMS setups (WordPress with two SEO plugins active, especially) output two Organization blocks. Parsers usually keep the first and toss the rest. If your hand-authored values schema comes second, it gets dropped. Check your page source for duplicate @type: Organization nodes.

Inconsistent entity naming. If your schema says "Acme Company", Wikidata says "Acme Co.", and LinkedIn says "Acme", knowledge graph systems have to guess which is canonical. Inconsistency lowers confidence in your record. Pick one form of your name and use it everywhere, schema included.

Over-nested structures. Schema can go many levels deep, but AI parsers that tokenize your structured data inside an HTML document run on token budgets. Deeply nested schema can get truncated. Keep your core brand identity properties at the top level of the Organization node.

How does schema markup interact with Google's AI Overviews and Gemini specifically?

Google has published more about how its AI uses structured data than any other provider, so this is the best-documented corner of the map [2].

AI Overviews (formerly SGE, now broadly live in the US) draw on three sources: Google's Knowledge Graph (which schema shapes heavily), live web retrieval, and the base LLM's training. For brand-specific queries, the Knowledge Graph is the most trusted source and the most schema-responsive one.

Google's entity extraction pipeline uses Organization schema to populate Knowledge Graph entries. A 2023 Semrush analysis found brands with complete Organization schema were 43% more likely to have a Knowledge Panel appear in Google Search, and Knowledge Panel presence tracks closely with appearing in Gemini responses about that brand [9]. The chain runs schema to Knowledge Graph to Knowledge Panel to Gemini entity retrieval.

For Gemini specifically, Google has confirmed that sameAs links to its own properties (Google Business Profile, YouTube channels) get extra weight in entity resolution. If you haven't linked your Organization schema to your Google Business Profile URL via sameAs, that's a quick win.

The Google AI Search landscape is moving fast, and the structured data signals that mattered in classic search are carrying forward into AI Overviews with more weight, not less. AI systems need reliable entity anchors to avoid hallucination.

The practical takeaway: a brand with a clear, consistent entity in Google's Knowledge Graph gets described by AI Overviews using that entity's properties. A brand without one gets a description synthesized from whatever prose the AI finds, which may or may not match your positioning.

What role does E-E-A-T play alongside schema for AI assistant citations?

E-E-A-T (Experience, Expertise, Authoritativeness, Trustworthiness) is Google's quality framework, and it shows up explicitly in how AI Overviews pick sources [8]. Schema is the technical layer. E-E-A-T is the credibility layer. You need both.

Schema tells AI systems what you claim. E-E-A-T signals tell them whether to believe you. A brand with strong Organization schema but no third-party corroboration gets cited less confidently than a brand with moderate schema and heavy credible coverage.

For values specifically, the E-E-A-T moves that pay off most: get your values claims cited by credible third-party publications, earn certifications that carry their own schema (B Corp, ISO standards, LEED), make your leadership team's expertise visible and linked through Person schema, and keep an author page for any content your brand publishes on values topics.

The full stack looks like this. Your About page has correct Organization schema with values properties filled in. Your CEO has a Person node linked to the Organization via worksFor. Your B Corp certification page carries EducationalOccupationalCredential schema. Your sustainability report has its own schema and cites its data sources. Each layer adds a credibility node AI retrieval systems can traverse.

Tools that audit your AI visibility can surface where the E-E-A-T and schema layers fall out of sync, which is often the real cause of a wrong AI description. Spawned's AI visibility audit checks for schema-content consistency as part of its brand positioning analysis, worth a look if you want a baseline for where you stand today.

Nobody has solid published data on the exact weighting of E-E-A-T versus schema in AI Overview selection. The closest reference is Google's own quality rater guidelines, which lay out the qualitative criteria without numerical weights [8].

How do you validate that your schema is actually influencing AI assistant outputs?

This is where most schema guides go quiet. Validating traditional rich results is easy: Google's Rich Results Test shows what parsed, and Search Console shows impressions. For AI assistant outputs, measurement gets harder.

The honest answer today is that no API tells you "your schema caused this AI response". What you can measure is a mix of leading indicators and outcome proxies.

Leading indicators (schema-level): Google's Rich Results Test [7] confirms your schema parses without errors. The Schema Markup Validator (validator.schema.org) confirms your types and properties are valid against the Schema.org spec. Google Search Console's Enhancements report flags structured data warnings at scale.

Outcome proxies (AI-level): Ask AI assistants brand-specific questions directly: "What does [Brand] stand for?", "Is [Brand] sustainable?", "What are [Brand]'s values?" Screenshot and date-stamp the responses. After a schema change, wait 4 to 8 weeks (Googlebot recrawl and Knowledge Graph update cycles), then re-query. Track whether specific claims from your schema show up.

Perplexity is the most useful platform for this because it shows source citations. When your About page starts appearing as a cited source for values queries, your schema is working at the retrieval level.

For a more systematic setup, tracking AI brand mentions against a baseline is what AI search visibility metrics tools are built for. The metrics that matter: citation frequency (how often an AI cites your owned pages), accuracy rate (how often the AI's description matches your positioning), and share of voice in AI responses for your core category queries.

One reality check: run the same brand queries across ChatGPT, Gemini, Claude, and Perplexity. You'll usually see meaningfully different descriptions. That divergence tells you which AI system's entity graph represents your brand most accurately, and where to spend your effort.

Is there a schema markup priority order for brand values positioning?

Working through this with limited engineering time? Here's the honest priority order, ranked by impact-to-effort.

Priority 1 (do this week): Fix any syntax errors in existing Organization schema. Add a description with a specific, verifiable values claim. Add sameAs pointing to at least two authoritative external profiles.

Priority 2 (do this month): Fill in knowsAbout with specific terms, not broad categories. Add an ethicsPolicy URL (make sure the target page exists and is crawlable). Add FAQPage schema to your About page with 4 to 6 real Q&As about your values.

Priority 3 (next quarter): Add ItemList schema for your core values on your About page. Add Person schema for key executives linked to the Organization. Add structured data to the sustainability or ethics policy page itself, beyond the organization node.

Priority 4 (ongoing): Keep schema consistent with content changes. Earn a new certification, add it via hasCredential. Update your values page, update the corresponding schema. Change brand messaging, update the description and slogan.

The biggest ROI for most brands is Priority 2. FAQPage schema is the most direct path to controlling AI answers to specific brand questions, and it needs no development work, just a JSON-LD block in the page head. A 2023 Semrush study found FAQ schema pages had 3.6x higher click-through rates in rich results than non-FAQ pages [9], and the same retrievability logic carries over to AI answer engines.

For the full AI SEO picture, schema is the foundation, not the whole building. Content quality, entity corroboration across the web, and technical crawlability all matter alongside it.

Sources

1. Schema.org, Organization type specification
2. Google Search Central, How AI Overviews work
3. Google Search Central, Structured Data General Guidelines
4. Perplexity AI, About page
5. Google Search Central, Structured Data for Knowledge Graph
6. Authoritas, AI Overview Citation Study 2024
7. Google, Rich Results Test tool
8. Google, Search Quality Evaluator Guidelines
9. Semrush, Structured Data and Rich Results Study 2023
10. Schema.org, Brand type specification
11. Schema.org, FAQPage type specification
12. Wikidata, main page