Conversational search is the shift from typing terse keywords ("running shoes flat feet") to asking a full, natural-language question ("which running shoes are good for a heavy runner with flat feet who mostly uses a treadmill?"), and often following up in a back-and-forth.
It's how people now use ChatGPT, Google's AI Mode, Gemini, and Perplexity. Queries are longer, more specific, and carry context a keyword search never did. Behind the scenes, the engine usually breaks one conversational question into several sub-queries. See query fan-out, and assembles an answer from many sources.
For brands, this changes the target. Thin pages tuned to one keyword win a shrinking slice of these questions; content that answers the whole intent. Uses, comparisons, who-it's-for, edge cases. Gets pulled into more of the answer. And because the answer often resolves in-chat, much of this traffic is zero-click.