I haven’t opened Google for a research question in three months. Not because I’m boycotting it — I just keep forgetting it exists. Perplexity answers my questions directly, with sources, and I never have to click through five SEO-optimized blog posts to find what I actually need.
This shift happened gradually, then suddenly. One day I realized my muscle memory had changed: research question → Perplexity, not research question → Google. That’s a big deal, and I don’t think I’m alone.
What Changed
Traditional search gives you a list of links. You click the first one, read half of it, realize it’s not what you need, go back, try the second link, hit a paywall, go back, try the third link, find the answer buried in paragraph 17 after scrolling past three newsletter signup popups and an autoplay video.
AI search gives you the answer. With citations. In 5 seconds.
The difference isn’t just speed (though that matters). It’s the elimination of friction between having a question and having an answer. No link-clicking, no popup-dismissing, no scroll-hunting. You ask, you receive.
The Players
Perplexity is the one I use daily, so let me be specific about why.
The citations are genuine. Every claim links to a real source. I can click through and verify anything that seems off. This isn’t ChatGPT confidently making things up — it’s more like having a research assistant who shows their work.
Pro Search mode is where the magic happens for complex questions. It thinks through the question, searches multiple sources, synthesizes the findings, and presents a nuanced answer. “Compare the privacy implications of running Llama locally versus using Claude’s API for a healthcare application” gets a detailed, multi-perspective response that would’ve taken me an hour of Google searching to assemble.
The follow-up capability is underrated. I ask a question, get an answer, then drill deeper: “What about HIPAA specifically?” Perplexity maintains context and builds on the previous answer. This conversational research flow is how I actually think about topics — I don’t know all my questions upfront.
Free tier is genuinely useful. Pro at $20/month is worth it for the enhanced model access.
Google AI Overviews appear at the top of Google results now, and they’re… fine. They answer straightforward factual questions quickly. But they feel bolted on rather than native. The overview sits on top of the old search results, and the whole experience is awkward — like a library that hired a receptionist books but kept all the bookshelves too.
For simple factual questions, they work. For complex, nuanced questions, they’re inadequate. And occasionally, they’re wrong in entertainingly obvious ways (remember the “eat rocks for minerals” incident?).
Copilot (Bing + GPT-4) is Microsoft’s entry. It’s better than Google AI Overviews but less polished than Perplexity. The best thing about it is the Edge browser integration — if you already use Edge, it’s right there. The image generation built in is a nice bonus.
Phind is specifically for developers, and it’s genuinely good at its niche. “How do I implement rate limiting in FastAPI with Redis?” gets a code-ready answer with explanation, not a link to Stack Overflow where the accepted answer is from 2019 and deprecated. If you write code, bookmark Phind.
When I Still Use Google
I’m not pretending AI search replaces everything. Here’s when I go back to traditional search:
Shopping. “Best noise-canceling headphones under $200” — I want to browse reviews, compare specs, read opinions. I don’t want a synthesized summary.
Navigation. “Open Gmail” or “Amazon customer service number.” I’m looking for a specific destination, not an answer.
Local stuff. “Italian restaurant near me open now.” Google Maps still wins here by a mile.
Current events. Breaking news, live scores, real-time information. AI search has latency — the web is usually fresher.
When I want diverse opinions. AI search synthesizes one answer. Sometimes I want to read five different perspectives and form my own view.
What This Means for Websites
If you run a website, this should worry you a little. AI search engines pull information from your content and present it directly to users — who then have less reason to click through to your site. Your content gets consumed without your site getting the visit.
The impact varies. Informational content (how-to guides, explainers, definitions) is most affected — that’s exactly what AI search summarizes well. Review content, opinion pieces, interactive tools, and e-commerce are less affected because users still need to visit the site for the full experience.
If you’re a content creator, the strategy shift is: create content that AI can’t replace. Original research, unique data, personal experiences, interactive tools, and community discussion are all things that AI search can cite but can’t replicate. “10 Ways to Improve Your Productivity” is summarizable. “I Tracked My Productivity for 365 Days — Here’s What Actually Worked” isn’t.
My Prediction
Within two years, most knowledge workers will use AI search as their primary research tool and traditional search for navigation and shopping. The tools will converge — Google is building AI into search, and Perplexity is adding more traditional search features.
The real winner will be whoever combines the best of both: AI-synthesized answers when you want them, traditional results when you want to browse, and the intelligence to know which one you need.
Right now, that’s Perplexity. But this race is far from over.
🕒 Last updated: · Originally published: March 15, 2026