For two decades, search has meant the same thing: type a query, scan a list of blue links, open five tabs, and synthesize the answer yourself. Google refined the algorithm, but the format stayed frozen in 1998. Then Perplexity arrived amongst the wave of AI tools with a different proposition: what if search engines actually answered questions instead of just pointing you toward answers?

I spend a lot of time researching productivity tools and how to use them with AI but it’s undeniable Google has been muscle memory for years. But after hearing a few friends rave about Perplexity's conversational search, I committed to using it as my default engine in an experiment.

The question I sought an answer to: whether Perplexity can deliver for real-world needs: work research, travel planning, tech troubleshooting, and the random curiosity spirals we all fall into. What I found? It fundamentally changed how I search, though not without exposing some significant cracks.

What Perplexity gets right about modern search

It treats context like a conversation

The most immediate shift was how Perplexity handles follow-up questions. When I searched "best project management tools for remote teams," it gave me a synthesized answer with clear citations. But here's where it diverged from traditional search: I could immediately ask "which ones integrate with Notion?" without re-explaining the entire context. Perplexity remembered we were discussing project management software and filtered its previous answer accordingly.

Google would have treated that follow-up as a completely new search, likely surfacing generic Notion integration articles. Microsoft Copilot does better here — it maintains conversational context within Edge — but it's embedded in Microsoft's ecosystem, making it feel more like a sidebar feature than a search replacement.

Google's Search Generative Experience (SGE) attempts similar context awareness, but it's still layered on top of traditional results, creating a hybrid experience that feels indecisive about its own purpose. Plus, with Perplexity launching its Comet browser, its adaptability for search will only grow multifold.

Where it saves actual time

The research assistant I didn't know I needed

The mental load reduction was tangible. For a piece on VPN services, I asked Perplexity to "compare NordVPN and Mullvad for privacy-focused users." It returned a structured breakdown: NordVPN's audited no-logs policy and server network size, Mullvad's anonymous account system and open-source commitment, and specific scenarios where each excels. Every claim was cited, usually a mix of privacy audit reports, Reddit discussions, and technical reviews.

I could have found this information through Google, but it would have required skimming four different reviews, cross-referencing features, and building my own mental model. Perplexity did the synthesis work upfront, and because it showed sources clearly, I could verify claims without starting from scratch. This helps offload the aggregation step while maintaining transparency.

For travel research, I tested "three-day itinerary for Kyoto in spring," thet type of query where Google usually surfaces SEO-optimized listicles stuffed with ads. Perplexity compiled suggestions from travel blogs, Reddit's Japan Travel community, and official tourism sites, organizing them into logical day-by-day plans. When I asked "which of these are accessible by subway," it filtered options immediately. The time saved wasn't just speed but also all the decision fatigue from comparing 15 browser tabs.

Where Perplexity hits the wall

AI has blindspots it won't admit

The problems emerged with niche queries. For example, when researching a question like “Does Shopify automatically collect sales tax for international digital product sales?” Perplexity confidently presented an answer but it was oversimplified. The sources it cited explained general sales tax collection rules for U.S. merchants, not the specific nuances of VAT, GST, or region-based digital tax obligations that vary by country. Google, with its sprawling index, surfaced an obscure Shopify community thread and a European Commission PDF clarifying how cross-border digital sales are treated under EU VAT rules.

This exposes Perplexity’s core limitation: it optimizes for confidence over comprehensiveness. When a query falls outside well-documented territory, it just synthesizes from whatever sources it can access, even if they’re tangential. Google’s blue links might feel antiquated, but at least they let you judge source relevance yourself. Perplexity makes that call for you, which works beautifully for mainstream topics but falters when you need depth.

Paywall gating created friction too. Multiple searches returned summaries based on paywalled articles I couldn't verify without subscriptions. Perplexity would cite a New York Times piece or academic paper, but actually reading the source required paying up. Google doesn't solve this either, but it doesn't pretend to. You know upfront when something's gated. With Perplexity, the synthesis made it feel like the information was accessible when it often wasn't.

When traditional search still wins

The index matters more than the interface

For highly specific software documentation, Google's exhaustive index is irreplaceable. Searching for "Claude API parameter streaming tokens" on Perplexity gave me a decent overview, but Google surfaced the exact documentation page in its first result. No synthesis needed, just the answer.

Shopping comparisons revealed another gap. Perplexity synthesizes reviews well enough, but it doesn't integrate shopping features like price tracking, inventory checks, or direct purchase links the way Google Shopping does. When I searched "best standing desk under $500," Perplexity gave thoughtful recommendations, but I still had to jump to Amazon or retailers to check availability. Google's ecosystem integration (Maps, Shopping, Gmail) creates connective tissue Perplexity can't replicate yet.

Google SGE occupies an awkward middle ground here. It adds AI-generated summaries above traditional results, but the summaries often feel redundant when blue links sit right below them. Microsoft Copilot integrates more naturally within Edge, but it's locked to that browser and Microsoft's broader services. Perplexity's strength is its independence — it's not trying to funnel you into an ecosystem, just answer your question.

The verdict: search finally has competition

It's not perfect, but it's legitimately different

After a week, I haven't switched back. Perplexity reduced my search time significantly for knowledge-based queries, and the conversational follow-ups made research feel less like tab juggling and more like thinking out loud with a knowledgeable partner. The transparency around sources builds trust in a way vague AI summaries don't. You can also use Perplexity beyond research purposes.

But it's not a universal replacement for search. For niche topics, shopping, and ecosystem-dependent tasks, Google's index and integrations still dominate. Perplexity works best as a complementary tool. It's become my default for exploratory research, explanations, and context-heavy queries, while Google remains the fallback for precision and breadth.

Search isn't going back to 1998. Perplexity proves the answer-first model works when done thoughtfully, and it's pushed Google and Microsoft to respond. That competition alone makes it worth adopting, even with its current limitations.