Entity optimization: one version of the facts about your brand
I bring everything known about your company into one consistent version: the name, the legal entity, people, products, dates and contacts. When sources agree, a model answers with confidence and names you. When they disagree, it picks whoever looks more consistent.
When do you need entity optimization?
When the problem is not a lack of content but contradictions between the sources that describe you.
You are confused with another company
A company with a similar name works in another industry, and the model blends the facts of the two into one answer.
Your data does not match
One address in directories, another on your site, a third in an old listing. The model cannot tell which source to trust.
Several brands or product lines
A group, sub-brands or a product range, and nobody outside can tell what belongs to whom.
The expert is not linked to the company
You are known personally, the company is known separately, and nothing on the web connects the two.
Outdated facts
The model quotes a former name, a CEO who left or a founding year you corrected on your site long ago.
No single source of truth
Every new profile or listing gets filled in from scratch – and each time a little differently.
What goes into
entity optimization?
Brand representation audit
- How AI answers describe your company for branded prompts
- Which sources the model relies on when it talks about you
- Completeness and accuracy of every statement
- Conflicts and ambiguities between sources
Entity and relationship map
- Brand, legal entity, trademarks and how they relate
- People: founders, executives, experts and their roles
- Services, products and product lines linked to their owner
- Locations, branches and divisions
Register of canonical facts
- The one correct spelling of the name in every form
- Dates, company details, contacts and service area
- Company descriptions in three lengths: short, medium and full
- Rules for using these facts in every new profile or listing
Brand fact sheet
- One document with every canonical fact
- A public facts page on your site
- A version for external platforms and journalists
- An update procedure for when the facts change
Your site’s pages
- About, Team and author pages reworked
- Product and service pages with correct attribution
- Facts aligned across every page of the site
- Explicit links between people, the company and services in the text
Structured data
- Organization or Person as the main entity of the site
- Service, Product, Offer and related types
- Verified sameAs links to external profiles
- A check that the markup matches the visible text
External sources
- Inventory of profiles, directories and knowledge bases: Google Business Profile, LinkedIn, Crunchbase, industry directories
- Sources ranked by their influence on AI answers
- A plan to bring each one in line with the register
- A re-check 30 days after the fixes
How does the project run?
Fact collection
A questionnaire and an interview: what is correct, what is outdated, what is disputed. 2 days.
Source audit
Where and how you are described, and what the model takes from where. 3–4 days.
Map and register
The entity map and the register of canonical facts. 3 days.
Fixes on your site
Pages, markup, sameAs links and the facts page. 3–4 days.
External sources
Corrections where there is access; a plan for the rest.
Re-measurement
Branded prompts checked again after 30 days.
What do you get?
The register of facts is a working document: every future profile and listing gets filled in from it.
Entity and relationship map
Brand, legal entity, people, products and services with explicit relationships between them.
Register of canonical facts
The one correct wording, dates, contacts and descriptions of the company.
Brand fact sheet
A short document for external platforms, partners and journalists.
Organization and Person markup
Structured data with verified sameAs links.
External source audit
Every profile, what is wrong in it, who can fix it and in what order.
Alignment plan
Next steps for the sources I have no direct access to.
Timeline and format
- Main work – 10–15 business days
- Corrections on external platforms – as fast as each platform responds
- Branded prompts re-checked after 30 days
- The register of facts comes in an editable format
What I need from you
- Confirmation of the facts: dates, company details, services and team
- Access to your directory and profile listings where you have it
- A decision on disputed points, such as which spelling is the main one
- CMS access for the changes on your site
Not included
- Creating a Google Knowledge Panel – no one can promise that
- Wikipedia articles and paid listings
- Fixing statements that damage your reputation – that is AI reputation management
How is the result measured?
This work is measured by how correct the mentions are, not by how many there are.
Fact accuracy
Share of branded answers that get your details right.
Source consistency
How many external profiles now carry the same data.
Brand distinction
Whether the model has stopped confusing you with similar companies.
What usually comes next?
Related reading
How to check your brand's visibility in ChatGPT, Gemini and Copilot
One prompt with your brand name proves very little. Here is the repeatable method I use: a fixed prompt set, repeat runs, a log of full answers, AI share of voice and what Bing and Google report.
Read →What is generative engine optimization (GEO), and how is it different from SEO and AEO?
GEO is the work of getting your business named and cited in AI answers. I explain how it relates to SEO and AEO, what to measure instead of rankings and what a realistic first 90 days look like.
Read →Questions about entity optimization
Will you get us a Google Knowledge Panel?
What about sources you have no access to?
Should we have a Wikidata entry?
We have several brands. Is that one project or several?
How soon will the model stop quoting the old data?
How is this different from filling in directory listings?
Discuss entity optimization
Send your company name. I will look at what AI systems say about you now and how far the data drifts between sources, and reply within one business day.
I reply within one business day. For a larger project, fill in the project brief.