The Brandologist
Big impacts

Entity crisis: AI doesn't see ghosts, here's how to show up

AI is honestly hustling pretty hard all day, every day. Between all of the players – Google AI Search (now the standard in Googling)/Gemini, ChatGPT, Claude, Perplexity, and whoever pops up tomorrow... there's plenty to target.

They are out there recommending businesses to strangers every single day. The question is whether yours is one of them. This is a practical guide to auditing your brand's AI visibility — what a Citation Audit actually involves, what entity authority means, and what the machines are looking for when they decide who's worth the name drop.

Kate Edwards
Kate Edwards
May 27, 2026

The Citation Audit

The most direct way to diagnose your entity authority is to ask the machines directly — and brace for whatever they say back.

A Citation Audit means querying AI engines with natural language prompts the way a real human would, then tracking whether your brand shows up, in what context, and with what confidence. Not ranked. Cited. Different thing entirely.

Here's how to run one without spiralling:

Prompt Testing — Build 15 to 20 strategic prompts across informational, comparative, and transactional intent. Think: "Who is the best [niche] specialist for [specific outcome]?" Run them through Gemini, ChatGPT, and Perplexity. Do it in an incognito window so your own search history isn't flattering you.

Presence Mapping — For each response, track: are you mentioned at all, where do you appear, and is the framing a genuine recommendation or just a passing mention? First citation carries disproportionate weight. A neutral mention is not the same as a trusted one. Both are better than silence, which is where most brands currently live.

Competitor Benchmarking — AI responses typically surface two to seven citations per query. If your competitors are filling those slots and you're not, you don't have a ranking problem. You have an existence problem. Note who's showing up, in what context, and start treating that as intelligence rather than a personal affront.

The Glossary: because it's the machines world, we just live in it...

A cheat sheet for the language your brand needs to be fluent in — whether you find it interesting or not.

Entity Linking: Connecting your brand to authoritative external sources (LinkedIn, Wikipedia, industry directories) so machines can confirm you are who you say you are. The digital equivalent of a reference check, except the referee is the entire indexed internet.

Author Entity Signals: your personal entity matters too. A byline strategy that links your name consistently across your site, LinkedIn, and any guest content tells the knowledge graph you're a real expert, not just a website.

sameAs: The specific schema property that does the entity linking. Tells every crawling system: yes, this LinkedIn profile, this Wikidata entry, and this website are all the same brand. Consistency across these is not optional.

Confidence Score: The internal weight an AI assigns to determine whether you're trustworthy enough to cite. Built slowly through consistent, corroborated signals. Eroded faster than you'd like. Treat it like a credit score for your brand's credibility.

Ground Truth: The source an AI trusts so completely it stops looking for alternatives. One entity per query tends to hold this position. The entire project is building toward this. Everything else is just contributing factors.

GEO (Generative Engine Optimisation): Optimising content to be retrieved and cited by AI engines rather than merely indexed by traditional search. A different discipline with different outcomes. Most brands haven't started. You now have no excuse.

Machine-Readable Blueprint: Your JSON-LD structured data. What you actually are, in a language machines speak fluently and your average human finds profoundly tedious. Both things are simultaneously true and neither changes how important it is.

Zero-Click Optimisation: Structuring your content so AI can lift a clean, accurate answer directly from your page without guessing. If the machine has to work to summarise you, it will summarise someone else instead. Make it easy or make peace with irrelevance.

NAP Consistency: (Name, Address, Phone) — boring name, non-negotiable practice. Every directory, every profile, identical. Machines cross-reference this obsessively.

Recency Signals: structuring content so AI can lift a clean, accurate answer directly from your page. If it has to work hard to summarise you, it won't.

Here's where to go:

A New Era for AI Search — Google I/O 2026

Google's own account of where search is heading — and it's not backwards. This is the announcement that confirmed AI is no longer a feature bolted onto search; it is search. Required reading if you want to understand what you're actually optimising for now.

5 New Ways to Explore the Web with Generative AI in Search

A practical breakdown of how AI Overviews and AI Mode are changing the way users actually interact with search results. Useful context for understanding why being cited matters more than being ranked.

AI Features and Your Website — Google Search Central

Straight from Google's developer documentation: how AI features work, what they pull from, and how to approach your content if you want a seat at the table. Not glamorous. Completely essential.

How Google Search Works — Google Search Central

The unsexy foundation everything else sits on. Crawling, indexing, ranking — if you don't understand the basics of how Google reads your site, the more advanced stuff is just noise.

SEO Best Practices for Ecommerce Sites — Google Search Central

Google's official guidance for product-based businesses navigating search. Structured data, product markup, and indexing best practices — the technical side of making sure your catalogue doesn't get ignored.

Web Stories — Google for Creators

Google's visual, tappable story format that lives on your own domain and surfaces in Search and Discover. An underused content format with solid discoverability potential, particularly for brands with strong visual identity.

Mangools — Five SEO tools in one account covering keyword research, rank tracking, SERP analysis, and backlink data. Accessible enough that you won't need a developer on speed dial to use it.

Writesonic — AI writing and content platform with built-in SEO features, useful for drafting blog content, meta copy, and product descriptions at scale without losing your mind.

Otterly AI — GEO Tools — A free suite of generative engine optimisation tools purpose-built for the AI search era, including a brand entity checker, prompt research, and citation tracking. Legitimately useful and not trying to upsell you a $500/month enterprise plan every thirty seconds.

Exploding Topics AI Visibility Checker — Semrush in a trench coat, yes, but the tool itself does what it says: checks how visible your brand is across AI search responses. Worth a run if you want a quick snapshot without setting up anything.