Answer Engine

Answer engines come in three families that behave quite differently. AI-native tools like Perplexity are built entirely around synthesized answers with visible sources. General assistants like ChatGPT, Claude and Gemini answer conversationally and search the web only sometimes. Hybrid features like Google's AI Overviews layer a generated summary on top of ordinary search results.

The distinction that matters most is where the information comes from. Some answers are drawn from training data — baked in months earlier and impossible to edit. Others come from live retrieval, fetched at the moment the question is asked. The fix differs completely: you cannot change what a model already learned, but you can influence what it finds when it looks.

Citation behaviour varies too. Perplexity shows sources for nearly every claim. ChatGPT cites only when it has actually searched. AI Overviews link a small number of pages beneath the summary. So being visible means something slightly different on each platform.

The practical consequence is that visibility has to be measured per engine. An aggregate number hides the case that matters most — being strong on one platform and absent from three.

An answer engine compresses the entire research phase into a single response. Where a results page gave ten brands a chance to be clicked, an answer names three and the reader stops. Being one of those three is worth more than almost any individual ranking position — and being none of them is a loss that never appears in your traffic reports.

Ask "best CRM for a 20-person sales team" on four engines and you will often get four different shortlists. Perplexity leans on whichever review sites ranked well that week. ChatGPT may answer from training data without searching at all. Gemini draws heavily on Google's index. A brand can be prominent on one platform and missing from the other three — a gap that stays invisible if you only ever check the one you personally use.