You type a question into the browser bar and something odd happens: instead of ten blue links, you get four paragraphs, a handful of numbered sources, and a short list of follow-up questions you didn’t ask. That’s Perplexity. Run the same query through Google and you get links, an AI Overview, a shopping carousel, three sponsored results and a map pack. Both answered. Only one felt like a conversation.
The perplexity vs google debate gets framed as a fight to the death. It isn’t. After roughly eighteen months of using both every working day, I’d argue they’ve quietly split the job between them, and the only real skill is knowing which one to hand a given question.
What Perplexity is actually built to do
Perplexity isn’t a search engine in the Google sense. It doesn’t hand you an index and let you sort it out. It runs your query across the web, pulls the relevant passages, and writes a synthesised answer with sources cited inline, usually somewhere between three and eight of them. Every claim has a number next to it. You can click straight through to the underlying page.
That design shines on questions that need assembly rather than retrieval. Something like “which London boroughs have the most EV charging points per capita, and how has that changed since 2020” is a nightmare in a traditional results page: you’d open six tabs, three of which would be council PDFs. Perplexity does it in about eight seconds and shows you where each figure came from. I’ve written before about what Perplexity genuinely does well and where it falls short, and the short version is that synthesis is the whole product.
The follow-up thread is the underrated part
Most people fixate on the answer. The thing that changed my behaviour is the thread. Ask a broad question, then say “narrow that to Scotland” or “what if I only care about 2024 onwards” and it just does it, keeping everything in context. Google makes you start over with a slightly reworded query and hope. That difference adds up over a working day.
Where Google still wins, and it isn’t close
Anything with a map, a price, or live inventory. Google knows which supermarket is open at 10pm on a Sunday, what a flight to Lisbon costs next Tuesday, and whether the shop two streets away has the cable you need in stock. Perplexity can describe those things. Google can actually settle them.
- Local and transactional queries — opening hours, directions, stock checks, reviews with photos.
- Navigational searches — when you half-remember a site’s name and just need the door.
- Breaking news — Google surfaces an hour-old report from a local outlet; Perplexity’s index often lags behind the minute-by-minute stuff.
- Anything you’d rather scan than read — sometimes ten headlines beat one confident paragraph.
The verification habit Google trains and Perplexity erodes
A results page is ugly, but it’s honest about its uncertainty. You see who’s saying what, you notice that two of the top results are affiliate spam, and you calibrate. Perplexity hands you a clean paragraph in a calm serif font, and the confidence of the formatting does a lot of persuasive work. It cites its sources, which helps enormously, but most people read the answer and skip the numbers. That’s the failure mode to watch for, especially on medical, legal or financial questions where a plausible-sounding synthesis can be quietly wrong.
The gap is closing from both directions
Google’s AI Overviews now do a decent impression of a Perplexity answer for simple factual queries, which has taken some of the shine off the newcomer. Meanwhile Perplexity has been adding shopping, local results and finance pages, edging toward the messy everything-engine it once defined itself against. Comparing Perplexity against Google, ChatGPT and Claude for everyday searching a year ago felt like comparing a scalpel to a Swiss Army knife. Now the blades look more similar, and the real differentiator is clutter.
Open Google for a commercial query and you’ll scroll past sponsored listings, a shopping strip and an AI summary before reaching an actual source. Perplexity’s free tier shows one sponsored follow-up question and gets out of the way. If ad-free research appeals to you, it’s worth reading about where ad-free engines like Andi fit alongside Perplexity and Google — the clutter question turns out to matter more than most people expect.
Cost and speed, in practical terms
Google is free and always will be, in the sense that you pay with attention. Perplexity’s free tier gives you unlimited quick searches and a small daily allowance of Pro searches, which run a deeper multi-step retrieval. Pro is around $20 a month and adds model choice, file uploads and far more Pro queries. For casual use, the free tier is genuinely enough. I paid for Pro for four months, dropped it, and only missed it on research-heavy weeks.
Speed is closer than the marketing suggests. A Perplexity quick search lands in three to five seconds. Google’s first paint is instant but the useful part — reading three results and working out which one answers you — often takes longer than the Perplexity answer did.
Which one to use for a specific job
- Debugging an unfamiliar error message — Perplexity. It’ll quote the two Stack Overflow threads that matter and explain the fix.
- Booking a haircut, a table or a train — Google, every time.
- Understanding a dense paper or a new regulation — Perplexity, then verify the key claims against the original.
- Buying anything with a warranty — Google Shopping, then a specialist review site.
- Comparing two options across six criteria — Perplexity, because it’ll build the comparison for you.
The workflow that stops you having to choose
My default is to start in Perplexity, because the first sixty seconds of any research task are about mapping the territory, and a cited synthesis beats scrolling. If the answer involves a place, a price or a purchase, I move to Google immediately. If a claim matters enough to repeat out loud, I click the citation and read the source, which is the step Perplexity makes easy and humans routinely skip.
For anything longer than a single question, I paste the Perplexity answer into a chat assistant and ask it to poke holes in the reasoning. That three-tool stack — researched answer, local verification, adversarial review — is what I’d recommend to anyone doing knowledge work. There’s a good breakdown of how ChatGPT and Perplexity divide that labour in a normal browser if you want to go further down the same path.
Google remains the place you go to act. Perplexity is the place you go to think. Keeping both pinned to the bookmarks bar, with a clear rule about which one gets which query, has saved me more time than any single subscription or upgrade I’ve made in the past two years. The engine that wins the perplexity vs google argument, in other words, is whichever one you reach for second.