Ask a generative engine who to hire for something and it does not consult one page. It assembles a picture of your business from every surface it can reach and checks whether those surfaces agree.
Agreement raises confidence. Contradiction lowers it. And low-confidence entities do not get recommended, they get skipped in favor of a business whose facts agree with themselves.
This is the most fixable problem in the entire discipline and the most commonly ignored.
What an Entity Actually Is
To a knowledge graph, your business is not a website. It is a node with attributes: a name, a category, a location, a set of services, a set of relationships to people and places, and a confidence score attached to each of those attributes.
Every surface mentioning the business is a vote on those attributes. Twelve surfaces saying the same phone number produces a high-confidence attribute. Nine saying one number and three saying another produces a low-confidence attribute, and the whole node gets treated more cautiously as a result.
The Surfaces That Have To Agree
- Website copy, especially contact, about, and pricing pages.
- Schema markup, particularly Organization or LocalBusiness.
- llms.txt at the site root.
- Google Business Profile.
- Bing Places.
- Apple Business Connect.
- Industry and local directories.
- Every social profile bio.
- YouTube channel description and about section.
- LinkedIn company page.
- Review platform profiles.
- Press mentions, guest posts, podcast bios, and anything a previous vendor created and forgot about.
That last one is where most of the damage lives. A listing created by an agency four years ago, with an old phone number and a service list that no longer describes the business, is still voting.
What Has To Match
Legal name, exactly. Not "UM Media" in one place and "Untethered Minds Media LLC" in another with no stated relationship between them. Pick a primary, use it consistently, and let schema carry the alternate as an alternateName.
Phone and address or service area. Same formatting, same number, same area. If the business is service-area rather than storefront, say so consistently everywhere.
Service list. The same lanes described the same way. A site listing five services and a directory listing two is a contradiction about what the business does.
Price ranges. If pricing is published anywhere, it has to be the same everywhere. This is the attribute that drifts most often, because prices change and listings do not.
Positioning sentence. The one-line description of what the business is. Reused verbatim, not rewritten per platform. Twelve slightly different descriptions is twelve slightly different businesses.
The sameAs Consolidation
Listing every controlled profile in the sameAs array of your Organization schema is what tells a search engine these accounts are one entity rather than several unrelated partial ones. Without it, a fragmented footprint stays fragmented, and the authority earned on each surface never consolidates into the node.
The Audit
Open a spreadsheet. One row per surface, one column per attribute: name, phone, address or service area, services, price range, positioning line, logo, URL. Fill it in by actually visiting each surface rather than from memory.
Every disagreement is a task. Fix the highest-authority surfaces first, then the long tail. Re-run it whenever anything changes and at least once a year.
It is unglamorous work with no visual output, and it is frequently the single highest-return item on an AEO or GEO list, because it is repairing damage rather than adding volume. The case for handling this inside one connected system is here.
Entity consistency is one of the first things repaired in a UM Media SEO / AEO / GEO Foundation. See what the Foundation covers, or the Cinematic Brand System.