/A necessary caveat
Nobody outside OpenAI knows the exact ranking mechanics of ChatGPT's recommendation behaviour. Below is a synthesis of publicly documented product behaviour, OpenAI's own guidance to publishers, and Digital Hive analysis of how NZ businesses actually appear in ChatGPT answers today. Treat it as a defensible working model, not a proprietary algorithm.
/Relevance
The first filter is basic: does this business match what the user asked for? Clear service statements, city or region, and specific offerings on your site make it easy for the model to identify you as a candidate. Vague marketing prose makes you harder to shortlist.
/Entity clarity
ChatGPT reasons about businesses as entities. The clearer your entity, name, legal name, location, category, services, the higher the chance the model can pick you up as a discrete recommendation instead of a fuzzy match.
/Third-party mentions
The model weights references to your brand across the web: press, blogs, directories, industry lists, partner sites. Several genuine references from reputable sources lift the confidence that you are a real, credible option worth citing.
/Citations from crawled sources
ChatGPT's browsing and retrieval components pull in structured citations. Pages that answer the question directly, cleanly formatted, well-linked, honest, are cited more readily than long marketing pages that hedge.
/Quality backlinks
Traditional backlinks still act as trust signals. Not every third-party mention is a backlink, but the same reputation logic applies: legitimate references from credible sources beat a hundred forgettable ones.
/Reviews
Public reviews across Google Business Profile and other trusted platforms feed the wider signal that a business is real and delivering. Fabricated reviews are increasingly detectable and hurt more than they help.
/Consistent business information
Name, Address, Phone (NAP) consistency across your own site and the wider web tightens the entity. Inconsistent details fragment you across ghost versions and dilute AI recommendations.
/Content clarity
Language models pull chunks. Definitions, lists, FAQs and comparison content are much easier to extract and cite than dense marketing copy. Structuring your site around the questions buyers actually ask is one of the highest-leverage AEO moves.
/Crawlability
Content locked behind JavaScript that render-only crawlers cannot access is invisible to a large slice of AI retrieval. Server-rendered content and semantic HTML remain baseline.
/Brand authority
Consistent brand mentions across the web build the sense that your company is a known quantity. A handful of legitimate press or industry references, over time, compound into a durable authority signal.
/Structured information
Schema markup helps machines confirm what you are, where and who you serve. It is not a magic switch, but it reduces friction and improves the odds of being included in an AI response.
/What NZ businesses can do this quarter
Audit your homepage for clear entity statements. Add schema. Tighten NAP across NZ directories. Publish two answer-first articles that genuinely help buyers. Ask for a handful of honest reviews. Consider whether one legitimate press mention is achievable this quarter. Small, consistent moves add up.
