Analysis
For 25 years, finding something online has worked the same way: type a few keywords, scan a page of blue links, click, and hope you guessed right. A new crop of products wants to throw that out. Instead of links, you ask a question in plain English and get a written answer with the sources cited underneath.
Three names are fighting over what comes next. Perplexity built an answer engine from scratch. OpenAI bolted search onto ChatGPT, the app hundreds of millions of people already open every day. And Google folded AI answers into the search box it has owned for a generation. Whoever gets this right won't just win a product category. They'll set the default for how most people look things up.
For an Australian business, the practical stakes are simpler than the hype suggests. If buyers start asking an assistant "who does X near me" instead of typing it into Google, the question of which engine answers, and whether your business shows up in that answer, becomes a real marketing problem rather than a futurist one.
Here's where each contender actually stands.
Perplexity: The AI-Native Pioneer
Perplexity was the first company to ship an AI search engine people genuinely wanted to use. Founded in August 2022, it never built a traditional search product at all (Perplexity AI, Wikipedia). Every query goes to an AI model that pulls together information from several sources and links each claim back to where it came from.
The strengths are easy to point at. Citation quality is the best in the field: claims are tied to specific sources, and those sources tend to be credible. The interface is clean and built for asking questions, without the clutter of a results page. Pro Search runs several searches at once and stitches the results into a fuller answer for harder questions, and Collections lets you save and organise research on a topic, updating it as new material appears (Perplexity AI, Wikipedia).
Reported user numbers vary by source and by how you count. Perplexity has cited figures around 34 to 45 million monthly active users on the core platform, and 100 million-plus across all its products (Perplexity AI Statistics 2026, DemandSage); an earlier draft of this piece put the figure at 65 million monthly users (up from 15 million a year prior), which we couldn't verify against current reporting. On funding, the company has raised well over $1.7 billion in total, with a 2026 valuation reported at roughly $22.6 billion after a Series E in January 2026 (Perplexity Funding & Investors 2026, Tracxn). Worth noting on the business model: Perplexity did experiment with advertising from late 2024, but reportedly pulled out of ads entirely by February 2026 over concerns it would erode user trust (Perplexity pulls the plug on ads, Campaign US).
ChatGPT Search: The Distribution Advantage
OpenAI's ChatGPT Search was announced in October 2024 and reached all users by February 2025 (ChatGPT, Wikipedia). Its edge is reach. ChatGPT serves an enormous audience: OpenAI reported 400 million weekly active users in February 2025, and by February 2026 reporting put that figure closer to 900 million (OpenAI now serves 400M users every week, TechCrunch). That's distribution Perplexity can only dream about. There's no new app to download and no behaviour to change; you flip search on inside the chat window you already use.
The pitch rests on two things: the models and the user base. Answer synthesis reportedly draws on OpenAI's newer GPT-5.5, released in May 2026, which the company says improves results when ChatGPT decides to search the web (Introducing GPT-5.5, OpenAI); whether it's the dedicated search-synthesis model isn't spelled out. Tying search to your conversation history means answers can lean on past chats, and following up ("what's the case against that?") feels smoother than the equivalent in Perplexity.
The gaps are just as visible. Citations aren't as tight as Perplexity's; sometimes the cited source doesn't clearly back the claim. The interface is built for chat, so it can feel busy when all you want is a fact. And it can lag on fast-moving topics, missing the newest information.
Gemini: The Knowledge Graph Advantage
Google's Gemini sits on top of the one asset nobody else has: Google's knowledge graph and live index. Where Perplexity and ChatGPT Search lean on third-party search APIs or limited crawls, Gemini taps the same data behind ordinary Google Search, with a quarter-century of ranking work behind it. Those figures, trillions of pages and 25 years of refinement, are fair descriptions of the scale rather than audited numbers (Google Search I/O 2026 updates, Google blog).
The upside is coverage and freshness. When Gemini answers, it's working from the most complete and current index going. Hooks into Maps, Shopping, Flights and Scholar add structured data the others can't match. And AI Overviews, the AI answer that sits above normal results, reaches around 2 billion monthly users, most of whom have never heard of Perplexity (Google Search I/O 2026 updates, Google blog).
The catch is delivery. Google's AI search has drawn criticism for uneven quality, the odd hallucination, and an experience that feels bolted onto traditional search rather than rebuilt around it. The deeper problem is incentive: Google makes its money from ad-driven search, and AI answers that keep people on the page cut into the clicks that pay the bills.