Generative AI
Generative AI is AI that makes new content, such as text, images or code, in answer to a request, instead of just sorting or labelling what exists.
Generative AI is AI that makes new content: text, images, audio, code. You ask for something and it writes or draws it, word by word or pixel by pixel. ChatGPT, Gemini, Perplexity and Copilot are all generative AI.
For marketers, it matters because it changed what a search result is. A search engine used to hand you a list of pages. A generative AI tool writes the answer itself, and decides along the way which brands to name.
How it differs from older AI
Older AI mostly sorted things. A spam filter decides whether an email is spam. A recommendation system picks which product to show you next. Each one chooses from what already exists.
A generative model writes something that didn't exist before. Ask "which coffee subscription is best for someone who likes light roasts?" and it produces a fresh paragraph, which might name Fernhill Coffee, might name Kestrel Roasters, and might change the next time you ask.
Why the answers vary
Generative models pick each next word from a range of likely ones. That's why two people asking the same question can get different answers, and why the same person can get a different answer an hour later.
So a single answer tells you little about how often a brand is named. You need many answers, collected over time, to see a pattern. That's why AnswerOrbit asks each prompt every day and bases every number on seven days of answers. How we count explains it.