Perplexity AI was founded in August 2022 by Aravind Srinivas, Denis Yarats, Johnny Ho, and Andy Konwinski — all researchers or engineers with backgrounds at OpenAI, Google, and other frontier AI labs. The founding thesis came from Srinivas's personal frustration with search: Google answered "how do I boil an egg" with ten blue links to recipe pages. The answer was always somewhere in those links — but the search engine never just told you. Perplexity would tell you.
The product definition was precise. Perplexity was not a chatbot like early ChatGPT — it was a search engine that happened to use language models to synthesise answers. The distinction mattered for two reasons: first, it meant Perplexity always cited sources (the answer came with links to the pages it had read, not just a generated response from training data); second, it meant the product was useful for queries where you wanted a current, factual answer, not a conversational exchange. "What is the current Fed interest rate?" "Summarise the latest research on metformin." "What actually happened at the Balenciaga ad scandal?" These were search queries, not chat prompts.
The mobile-first approach was deliberate and differentiating. While competitors built primarily for web, Perplexity invested heavily in its iOS and Android apps from early in the product's life. The mobile experience was designed around the voice query — asking a question out loud and receiving a synthesised answer with sources was qualitatively better than typing into a search bar and scrolling through results. By 2024, mobile accounted for a majority of Perplexity's query volume.