// Entity & Brand Optimisation

Consistency Earns Trust. Markup Doesn’t.

Five minutes on what actually gets your brand cited with confidence in AI search, and what’s leftover SEO folklore that never worked the way it was sold.

Talk to us about your brand’s footprint

The Problem

Everyone Is Telling a Slightly Different Story About You

Your brand’s facts live in a dozen places at once: your own site, Wikipedia, Wikidata, LinkedIn, Crunchbase, industry directories, review platforms. When those sources agree, a language model treats your details as settled and states them with confidence. When they don’t, independently published research on how these systems handle conflicting sources is consistent: the model defaults to majority-vote guessing, blends the versions into something nobody actually said, or drops the disputed detail entirely rather than risk stating it wrong.

None of those outcomes help you. A guess can be wrong. A blend can misrepresent you. And silence, the model simply omitting you, is its own kind of erasure.

The Folklore

Why NAP Consistency and Schema Markup Alone Don’t Do What They’re Sold to Do

Two of the most repeated tactics in brand and entity SEO were built for an older, different system: Google’s traditional Knowledge Graph. Neither has a documented path into how a generative model actually decides what to cite.

NAP Consistency

Matching your name, address, and phone number across directories helps you get pulled into a structured local database, like Google Business Profile. It has no separate, documented effect inside a language model itself. That’s confirmed directly against Google, OpenAI, and Anthropic’s own platform documentation, not inferred.

Schema.org Markup

Google’s own engineering documentation states plainly that no special schema.org markup is required to appear in AI Overviews or AI Mode. Independent testing backs it up: Ahrefs added JSON-LD schema to nearly 1,900 pages and found no meaningful citation lift on Google AI Overviews, AI Mode, or ChatGPT. Markup still earns rich snippets in classic search. It just isn’t the AI-citation lever it’s sold as.

What The Evidence Shows

The Evidence Here Is More Contested Than Most GEO Advice Admits

A widely cited 2023 benchmark from Princeton and IIT Delhi found that citation-rich, quote-heavy content produced up to a 41% visibility lift on a metric called Position-Adjusted Word Count. That headline number came from a simulated benchmark using GPT-3.5-turbo as a stand-in for a real search engine, not a test against production AI search. The same paper's own smaller live test against real Perplexity, run that same year, told a weaker, messier story: quotation additions gained roughly 22%, not 30 to 40%, and keyword stuffing performed 10% worse live despite looking neutral in simulation.

A 2026 study went further and tested the same tactics directly against today's production models, GPT-4o-mini, Gemini, and Qwen-plus, and found they don't reliably improve citation visibility over doing nothing at all, and in several cases perform worse. The 2023 advice to simply add quotes and statistics is still circulating as current best practice. The more recent evidence says it's contested, not settled.

What holds up more consistently: restructuring facts into clear, human-readable content, definitions, tables, and relationships stated directly in the visible page, not buried in a script tag, showed a real accuracy lift in a separate 2026 study on structured entity pages. Worth knowing before you weigh that finding too heavily: its authors work for a company that sells structured-data tooling, so treat it as one credible data point, not independent confirmation.

Where We Fit

Where Entity and Brand Optimisation Fits

We align the facts about your brand across the places that still carry real, separate weight: Google Business Profile and Knowledge Panel data for local and traditional search, and Wikipedia, Wikidata, LinkedIn, and Crunchbase for the sources language models actually draw consensus from. Then we restructure how those facts appear in your own visible content, so a model finds one clear, quotable, internally consistent answer instead of stitching one together from fragments.

What we don’t do is sell you a markup checklist and call it AI optimisation. The research doesn’t support that claim, and we’re not going to write copy that pretends otherwise.

What Changes

One Fix, Three Audiences

Closing the consistency gap plays out in three places at once, because three different audiences are reading the same set of facts.

01 / The Buyer

A consistent story about who you are removes doubt before it forms. They don’t have to fact-check you across five open tabs to feel safe choosing you.

02 / The Search Engine

Consistent facts across authoritative sources are exactly what traditional Knowledge Panel and local search systems are built to reward.

03 / The AI Agent

When your facts agree everywhere, a model doesn’t have to guess, blend, or drop you. It cites you.

This is one piece of what Answerability audits, alongside content architecture, technical access, and performance measurement.

See the Full Answerability Approach

Questions

Frequently Asked

What does Entity and Brand Optimisation actually check?

Whether the facts about your brand agree across your own site and the outside sources language models actually draw consensus from, and whether your own content states those facts clearly enough, in visible text, not just markup, for a model to extract and quote with confidence.

Will this get schema markup added everywhere?

We use it where it still earns something real: traditional rich results and shopping or product feeds. We don’t sell it to you as an AI-citation lever, because the research, including Google’s own documentation, says it isn’t one on its own.

Is this the same as the Answerability audit?

No. Our Answerability product page audit scores individual product and category pages against the specific questions a buyer needs answered before purchase. Entity and Brand Optimisation works at the brand level, across every place your brand’s facts appear, not one page at a time. Many businesses eventually need both.

How does this fit with the rest of Answerability?

Entity and Brand Optimisation assumes your site is already technically accessible to AI crawlers, and that your content actually answers a real buyer's question. If access is the open question, start with Technical AI Search Optimisation. If it's whether your content itself is built around real buyer questions rather than a keyword list, that's Content Architecture for AI Citation. And if you need to know whether any of this is actually working, that's AI Search Performance Measurement.

When is this not the right fit?

If your brand doesn't have a public footprint yet, just one site, nothing else, there's nothing yet to make consistent. Past that, we honestly can't think of a legitimate reason a real business wouldn't want this. Staying hard to pin down, letting your details drift and disagree across the web, is mostly a pattern we see from churn-and-burn operations and sites that don't want to be easily checked up on. If that's not you, this is a fit. If your actual problem is a specific product page not answering a buyer's question rather than your brand's story being inconsistent, that's a page-level fix. Our Answerability audit handles that instead.

Find Out Where Your Brand’s Story Disagrees With Itself

We’ll show you where your facts conflict across the sources that matter, and what a language model does with that conflict right now.