Guide
How AI Chooses Which Businesses to Recommend
AI platforms do not rank pages. They interpret organisations. When someone asks "recommend a [service] near me," the AI builds a probabilistic model of every business it knows about in that category and selects the one it has the most confidence in.
That confidence comes from five things: can the AI clearly describe what the business does? Does it specialise in something specific? Is the website structured so AI can extract information reliably? Do independent sources confirm what the website says? Has the business been consistent over time?
The Key Insight
Brands with strong entity signals across 5+ platforms are 4x more likely to be recommended. It is not about having the best website. It is about having the most consistent, verifiable presence across the most trusted sources.
A small specialist firm with clear identity, subject authority in a defined niche, good technical foundations, and a growing ecosystem of independent references will often outperform a large general business with a poor entity signal structure.
Each Platform Sources Differently
ChatGPT – Bing index + training data. Submit to Bing or be invisible.
Gemini – Google's index + Knowledge Graph + Google-owned platforms.
Perplexity – real-time open web crawler. Explicitly cites sources.
Claude – indexed web + training data. Values consistent entity definition above all.
Copilot – Bing + LinkedIn. LinkedIn weighted more than any other platform.
Google AI – Google's own infrastructure. Content on Google platforms is trusted immediately.
Despite sourcing differently, every platform evaluates the same foundational signals. Get the baseline right and every platform benefits simultaneously.