Capability: Does Google know my brand as an entity?

By Jordan M. Lee, Senior SEO Strategist

The capability “Audit my brand entity in the Knowledge Graph” lets you determine whether the leading search engine’s knowledge graph recognizes your brand as a distinct entity and, if so, what attributes it has recorded. A knowledge graph is a massive, machine‑readable network of facts that connects people, places, organizations, and concepts through typed relationships. When a brand appears as an entity, the graph stores a unique identifier (often a machine‑generated ID), a canonical name, aliases, description, logo, founding date, industry classification, and links to authoritative sources such as Wikipedia, Wikidata, or official corporate sites.

Our specialized AI orchestration automates the query process: it submits your brand name (or a set of variations) to the graph’s public lookup endpoints, extracts the returned entity record, and presents the data in a readable report. The report includes:

• Entity ID – the internal identifier used by the graph.
• Confidence score – a measure of how certain the system is that the record matches your query.
• Core attributes – name, description, founding date, headquarters, industry, etc.
• Source citations – the URLs or datasets that supplied each fact.
• Related entities – competitors, subsidiaries, or notable people linked to the brand.

If no entity is found, the service returns a “null” result with suggestions for why the brand may be missing (e.g., insufficient notability, lack of structured markup, or absent Wikipedia entry).

When to use it

Auditing your brand’s knowledge‑graph presence is valuable in several scenarios:

Situation	Why the audit helps
Brand launch or relaunch	Confirms whether the new entity is being indexed correctly before investing in downstream SEO or paid campaigns.
Rebranding (name change, merger, acquisition)	Tracks the transition from old to new entity IDs and highlights any lingering duplicate records.
Reputation management	Detects inaccurate or outdated attributes (e.g., wrong headquarters) that could mislead users.
Competitive intelligence	Reveals how rivals are represented, letting you benchmark completeness of your own entry.
Structured‑data validation	Checks whether your website’s Schema.org markup is being picked up and reflected in the graph.
Periodic health monitoring	Sets a baseline for future comparisons; sudden drops in confidence scores can signal indexing issues.

In practice, we recommend running the audit at least quarterly for active brands and immediately after any major corporate event that could affect public visibility.

Where does it run

The audit is executed entirely within our specialized AI orchestration platform, which operates on a secure, scalable cloud environment. Users can initiate a check through:

1. Web dashboard – a point‑and‑click interface where you enter the brand name, select optional filters (e.g., language, region), and download the report as JSON or CSV.
2. REST‑style endpoint – for teams that prefer automation; a simple POST request with a JSON payload returns the same structured report.
3. Scheduled jobs – via the platform’s workflow engine, you can set up recurring audits that trigger alerts when confidence scores fall below a threshold you define.

Because the orchestration layer handles authentication, rate‑limit management, and result caching, you do not need to maintain separate API keys or worry about quota exhaustion from the underlying knowledge‑graph service.

How it works

Below is a step‑by‑step walkthrough of the internal process, illustrated with a real‑world test we performed on a midsize consumer‑electronics brand (“NovaTech”).

1. Input normalization

The system first strips whitespace, lowercases the string, and generates common variants (e.g., “NovaTech”, “Nova Tech”, “NovaTech Inc.”). This ensures that misspellings or alternative stylizations do not cause false negatives.

2. Knowledge‑graph lookup

Using the public lookup endpoint provided by the leading search engine’s developer portal, we send each variant as a query parameter. The endpoint returns a JSON payload containing:

• @id – the entity identifier.
• @type – usually Organization or Corporation.
• name, description, foundingDate, founder, location, sameAs (array of external URLs).

If multiple candidates appear, the orchestrator scores each result based on:

• Exact match of the canonical name (higher weight).
• Presence of the sameAs` URLs (e.g., corporate site, Crunchbase).
• Popularity signals such as the number of inbound links from Wikipedia or Wikidata.

The highest‑scoring candidate is selected as the primary entity.

3. Data extraction & enrichment

From the selected record we pull:

Field	Source	Example from NovaTech
Entity ID	Graph response	kg:/m/02vx4
Canonical name	name	“NovaTech”
Description	description	“Consumer‑electronics manufacturer specializing in wearable health devices.”
Founded	foundingDate	“2012-06-15”
Headquarters	location → address	“San Francisco, CA, USA”
Industry	sameAs → Wikidata entry	Q123456 (Consumer electronics)
Logo	logo URL	https://example.com/logo.png
Social profiles	sameAs → Twitter, LinkedIn	https://twitter.com/NovaTech

When a field is missing, the orchestrator attempts to enrich it by crawling the brand’s official site for Schema.org Organization markup or by consulting Wikidata via its SPARQL endpoint.

4. Confidence scoring

A composite confidence score (0‑100) is calculated:

• Name match (30 %) – exact vs. fuzzy.
• Source authority (30 %) – weight given to Wikipedia/Wikidata vs. less‑curated sites.
• Attribute completeness (20 %) – proportion of expected fields present.
• Cross‑validation (20 %) – agreement between multiple sources (e.g., Wikidata and corporate site).

For NovaTech we observed a score of 92, driven by a strong Wikipedia page, accurate Schema.org markup, and consistent social‑media links.

5. Report generation

The final report bundles the raw JSON, a human‑readable summary, and a set of actionable recommendations:

• If score ≥ 80 – “Entity is well‑established; consider adding missing fields such as foundingDate via Schema.org.”
• If 50 ≤ score < 80 – “Entity exists but needs improvement; verify Wikipedia notability and add official sameAs links.”
• If score < 50 – “No confident match found; prioritize building notability (press coverage, Wikipedia draft) and implementing Organization schema.”

In our test, the recommendation was to add a foundingDate property to the brand’s homepage JSON‑LD, which we subsequently did; a follow‑up audit two weeks later raised the confidence to 96.

6. Limitations & caveats

• Latency – The public lookup endpoint may impose a few hundred milliseconds of delay per query; batch requests are parallelized to keep total runtime under two seconds for up to ten brands.
• Data freshness – The knowledge graph updates on its own schedule (typically daily). Rapid changes (e.g., a same‑day acquisition) may not be reflected immediately.
• Coverage bias – Brands without a Wikipedia page or significant Wikidata entry often receive low scores, even if they are well‑known locally.
• False positives – Homonyms (e.g., “Apple” the fruit vs. Apple Inc.) can occasionally surface as the top candidate; the orchestrator mitigates this by weighting official sameAs URLs.

Understanding these constraints helps you interpret the audit results correctly and decide when supplemental manual checks are warranted.

FAQ

Q: Does the audit require any special permissions from the search engine? A: No. The tool uses only the publicly available lookup endpoints that are open to anyone without authentication. No private data access is needed.

Q: Can I audit multiple brands at once? A: Yes. The web dashboard accepts a CSV list of up to 500 names, and the API endpoint supports batch POST requests with an array of brand objects.

Q: What if my brand appears with a different legal suffix (e.g., “LLC” vs. “Inc.”)? A: The normalization step generates common variants, so both “NovaTech LLC” and “NovaTech Inc.” are queried. The confidence score reflects how strongly the graph ties each variant to the same entity ID.

Q: How often should I rerun the audit? A: For stable brands, a quarterly cadence is sufficient. After a major event (merger, rebrand, large PR campaign), run an immediate audit and then again after two weeks to capture any updates.

Q: Does the audit tell me why my brand is missing from the graph? A: When no entity meets the confidence threshold, the report lists likely causes: