How to Rank in AI Chatbots, Not Just Google

“Ranking” in AI chatbots is a category error — and understanding why it’s a category error is the first step to actually achieving what you’re trying to achieve.

Google ranking is a position. Position one means you appear above position two. There is a direct, measurable relationship between position and traffic.

AI chatbot visibility is different. Your business either appears in the answer, or it doesn’t. When it appears, it may appear as the primary recommendation, as one of several, or as contextual detail. The mechanic is not positional — it is citational.

The question to ask is not “how do I rank in ChatGPT?” It is: “how do I become the entity that AI systems cite when someone asks about my category?”

The citation mechanic across different engines

ChatGPT (GPT-4o, GPT-4): Draws primarily from training data with a knowledge cutoff. Cites businesses that have strong entity presence in the structured sources its training data included. The intervention is building better entity presence in those sources — Wikidata, trade press, structured directories — so future training runs include stronger signals about your business.

Claude: Similar to ChatGPT in architecture. Trained on a corpus that includes structured web data. Entity coherence in authoritative sources is the primary signal.

Perplexity: Retrieves live web content for most queries. The citation mechanic here is closer to traditional SEO — your pages need to appear in retrieval results, and then be structured well enough that Perplexity extracts and cites them. More responsive to changes than model-based engines.

Google AI Overviews: Powered by Gemini with access to Google’s full index. Weights Knowledge Graph entity strength, schema markup, and E-E-A-T signals heavily. The most directly influenced by traditional entity work because Google is the Knowledge Graph operator.

Gemini standalone: Similar to Google AI Overviews in terms of what it weights. Strong Knowledge Graph presence is the primary lever.

Each engine has a slightly different mechanic, but all of them share a common foundation: entity coherence and authoritative corroboration.

The entity coherence framework

Entity coherence means AI systems can answer the question “what is [your business]?” with a consistent, accurate, well-supported description. This requires:

A canonical identity. A single, consistent answer to: what is your business called, what category does it belong to, what does it do, where does it operate, who does it serve. Applied consistently across every source on the web.

Structured machine-readable data. Schema.org markup on your website that describes your entity attributes directly. This is not optional — it is how your website connects to your entity record in the Knowledge Graph.

Wikidata entry. The structured, machine-readable layer that Wikipedia is built on. Every business can have one. It directly feeds the entity layer that AI systems reference, and it is one of the most under-utilised tools available to any business working on AI visibility.

Authoritative corroboration. At least three to five authoritative external sources — industry press, accreditation bodies, structured directories — that independently describe your business in consistent terms. Each corroboration adds weight to your entity claim.

Consistent NAP. Name, Address, Phone (for local businesses). Inconsistency across directories and listings creates contradictory entity signals that AI systems resolve by… not mentioning you, or getting you wrong.

The content layer

Structured entity work makes your business citable. Content structured for AI extraction makes your specific claims citable.

The difference is this: entity work makes AI systems know who you are. Content work makes them know what you’re saying — and makes your specific insights, expertise, and services appear in answers.

Content optimised for AI citation shares characteristics:

• Direct, extractable answers in the first paragraph
• Specific, attributable claims (not vague generalities)
• Structured formats (Q&A, numbered steps, comparison tables) that AI systems can parse
• Clear authorship attribution with verifiable expertise signals

Measuring it

Traditional rank-tracking tools do not measure AI chatbot visibility. The measurement approach is different:

• Manual query testing across four to five key engines
• Recording which queries produce citations of your business
• Tracking the accuracy and completeness of those citations
• Monitoring competitors’ citation rates for the same queries

This is what a citation tracker does — and it is the measurement infrastructure needed to know whether entity work is actually moving the needle.

Your baseline today, however poor, is the reference point. Improvement from that baseline, over time, is the return on entity investment.