Entity Signals for Company Websites: About, Team, Contact, and Case Studies

The short version #

Before a search engine or an AI summary decides whether to cite your company, it tries to confirm you exist as an entity. It cross-checks the company name, address, phone, founders, social profiles, and named clients across your site and the open web, looking for consistency. That is what entity signals are. If your About page says "professional team," your team page is empty, and your case studies are anonymous logo walls, no amount of blog content saves you. The AI quotes a competitor whose pages have higher information density. This article walks through how to make your About, team, contact, and case pages function as verifiable entity signals instead of decoration. The audit table at the end is the version to print before your next content meeting.

We ran a small experiment last quarter. Same niche, five companies, three AI search tools. We asked each, "Which Chinese supplier can do X-type OEM work?" The most-cited company wasn't the largest or highest-ranked. It was the one whose About page named the founder with a LinkedIn link, whose case studies told real client stories, and whose contact page listed a real email and a WhatsApp number alongside the form. Their product writing was unremarkable. The AI could just verify them as a real company from five different angles.

This piece is for the person opening their own About, case, and contact pages right now and asking, "Is this enough?"

1. What an entity is #

Search engines no longer just match keywords. They maintain a knowledge graph in which every real company, person, and product is a node connected by verifiable relationships. Google Search Central's guidance on AI features is explicit about this. They want authors, organizations, and sources clearly identifiable.

AI search products (Perplexity, ChatGPT search, Google AI Overviews) follow the same logic. The 2023 paper Generative Engine Optimization (arXiv:2311.09735) found that cited pages in generative answers share three traits: clear stance, verifiable evidence, identifiable source. That last one is mostly entity signal.

For a small or mid-sized company, this isn't abstract. It comes down to five things:

• The company name is written the same way on your site, LinkedIn, WhatsApp Business, X, and industry directories.
• Your About page shows specific founders or team members, not stock photos.
• Your case studies show real client names (with permission) and verifiable outcomes.
• Your contact page has a real human channel, not just a form into a black hole.
• Your structured data and your visible page content tell the same story.

Miss one and your citation rate drops. Miss three and you basically don't get cited.

2. About page #

The default Chinese B2B About page is a hero photo, a CEO quote, and a string of adjectives: "professional," "leading," "one-stop." A Western buyer reads that and assumes the company won't reply to email. An AI summary reads it and sees almost no information density worth quoting.

A working About page answers four questions in plain language:

• Who you are: legal entity name (in English and the local language), country and city of registration, year founded. These are the hard facts an AI uses to triangulate you against the open web.
• Who runs it: at least two or three named team members with their roles and short professional backgrounds. Link to LinkedIn if you can.
• Who you serve: industries, company sizes, geographies. "Global clients" doesn't count.
• How to reach you: at minimum a business email and a real office address. A P.O. box doesn't count as an address.

Founders sometimes push back. "We don't want to put faces on the site." Fine, but understand the trade-off. If you won't show people, you have to over-deliver on the legal entity, registration details, named clients, and external anchors. Something has to do the verification work.

While you are auditing, line your About content up against the Overseas Social Profile Consistency Checklist. Same company name, same one-line description, same contact details across every channel. Three different "About" blurbs across three channels actively breaks entity consistency.

3. Team page #

The team page is the page Chinese sites skip most often, and the page Western buyers read most carefully. We've seen team pages that are nothing more than a group photo and three job titles, with no names. That contributes almost nothing to entity signals.

A useful team page has, for each person:

• Name, written consistently. Pick English name or pinyin and stick with it across LinkedIn, X, and email signatures.
• Role plus one sentence of what they actually do. "COO" is a label. "Owns overseas channel partnerships, focused on Europe and the Middle East" is information.
• A LinkedIn link, if available. It is the cheapest external anchor an entity can have.
• A real photo. If you genuinely cannot, leave the avatar off rather than use AI-generated headshots. Tools are getting better at flagging those.

You do not need 80 people on the page. Five to eight people who represent the company externally is plenty. When an AI verifies the entity, it walks the chain name → LinkedIn → your domain. The shorter and more complete that chain, the higher the trust score.

4. Case studies #

Case studies are the most under-used asset on most company sites. A single well-written case page can carry the entity load on its own. Client name, industry, problem, action, outcome, every line verifiable.

What a usable case includes:

• Client identity: company name (with permission), industry, region, rough size. Even under NDA, "a European appliance brand, around €500M annual revenue" is far stronger than "a leading client."
• Real problem: one or two sentences naming the actual issue. "They wanted digital transformation" is not a problem. "Their existing site loaded in 8 seconds from Germany and the German pages were machine-translated" is a problem.
• What you did: not "delivered an end-to-end solution." The specific actions. Migrated hosting to Frankfurt, rewrote German service pages with a native writer, configured hreflang.
• Verifiable outcome: relative numbers are fine. "Inquiries grew from 4 per month to 11 per month" is worth a hundred times more than "significantly improved."

We've written separately about case-study SEO mechanics in Image and Case Study SEO: alt text, internal linking, structured data. The point here is simpler. A case study isn't a marketing piece, it is evidence. When an AI generates "which company should I use for X?", what it pulls from is exactly the named clients and concrete numbers on case pages.

Anti-pattern: a case study that ends with "the client expressed great satisfaction with our service," no client name, stock photo. That page has near-zero entity density. Neither an AI nor a buyer will treat it as evidence.

5. Contact page #

The implicit question an AI is asking when it reads your contact page is whether a real human can reach this company and get a reply within 24 hours. If the answer looks like no, the company gets weighted down as possibly inactive.

What belongs on a contact page:

• A real office address. If you are remote, at least the registered address.
• A business email on your own domain, not Gmail or 163.
• A phone number or a WhatsApp Business number, with country code.
• Working hours, with time zone.
• A form with no more than five fields.

What does not belong:

• A form with eight required fields including "company size" and "annual revenue."
• A "we'll get back to you" message tied to an email nobody monitors.
• An address that stops at the city.
• A bot widget that prevents the visitor from sending a plain email.

We've seen Chinese companies whose entire contact page is "scan to add WeChat." That's efficient domestically. Overseas it is a closed door. Buyers don't scan unknown QR codes to add personal IM accounts before they've introduced themselves.

6. Structured data #

The five sections above are the visible layer. There is a second layer for search engines and AI: structured data, or schema. Its job is to declare the same facts in a machine-readable format so crawlers don't have to guess.

The ones that matter most:

• Organization schema on the homepage or About page: company name, logo, address, contact details, social profile URLs.
• Person schema on team member pages: name, role, employer, LinkedIn URL.
• CaseStudy or Article schema on case pages: title, author, publish date, client industry.
• FAQPage schema on FAQ blocks or dedicated FAQ pages.

For field-level templates, see Schema Markup for Service Websites. One trap to call out is schema that disagrees with the visible page content. Schema says the office is in Shanghai, the page says Shenzhen. Schema lists a founder name that appears nowhere on the site. When AI catches this, it discounts the whole site, which is worse than having no schema at all.

Simple rule: get the visible facts right and consistent first, then let schema declare the same facts to machines. Schema is a declaration, not a patch.

7. Trust audit #

Run this self-audit on your own site. Print it, walk through it, score honestly:

If four or more rows are blank, About and Cases are the two cheapest places to start. Fixing those two does more for entity signals than publishing 50 new blog posts.

A few things that look like entity work but aren't #

Common moves that look like they strengthen entity signals but actually erode trust:

• AI-generated team members: stock-style headshots, fabricated English names, generic titles. AI verifiers cross-check; when they find no external anchor for these "people," they discount the entire site.
• Keyword-stuffed About pages: "export, OEM, ODM, one-stop, professional, world-class" packed into the opening paragraph. Reads like nobody specific. Entity density goes negative.
• Logo walls without case studies: thirty client logos and zero stories. AI treats this as decoration, not evidence.
• Conflicting "founded in" dates: 2015 in one place, 2018 in another, 2020 in the structured data. Internal inconsistency is worse than a missing fact.

GEO is still moving, and we will not pretend to know exactly how the next AI Overview update will rank pages. What is clear is that verifiability beats polish. A plain About page where every claim checks out gets cited more often than a glossy one made of adjectives, which lines up with what Google has been saying for years in its Helpful Content guidance: write content from which real users get real information.

FAQ #

Do we have to put team photos on the site? #

Not strictly. But you need substitute evidence. If you skip team photos, over-invest in your legal entity name, registered address, LinkedIn company page, and detailed case studies. We've seen companies with no faces get cited at high rates because every other anchor is rock solid.

What if case clients won't let us use their name? #

Use industry, scale, and region. "A German Tier-1 automotive supplier, around €1B in revenue" carries far more entity weight than "a well-known client." Fully anonymous case studies contribute almost nothing.

Our company is three people. Will the team page look thin? #

Three real people beat thirty fictional ones. Write the names, roles, professional backgrounds, and LinkedIn profiles in full. Add three to five real case studies. That's a complete entity signal, not a thin one.

How fast can we see results in AI summaries? #

Usually four to twelve weeks. AI search indexes faster than traditional SERPs, but it isn't instant. In our experience, after a substantial rewrite of an About or case page, third-party AI tools start describing the company differently within one to two months.

Next step #

If you are unsure whether your About, team, and case pages are pulling their weight in AI citations, bring your domain, target markets, and social profiles for an enterprise AI and SEO/GEO audit. We will walk the trust audit table with you and tell you which two or three items are dragging the rest down, and which can wait. If schema, hreflang, or UTM are still fuzzy, our overseas website glossary is the fastest place to start.