You’ve likely heard the name. Perplexity.
It’s the AI-powered search engine that boldly claims it’s here to challenge Google.
But beyond the hype, one question remains:
Does anyone actually use it?
Surprisingly, yes—and in growing numbers.
Launched in late 2022, Perplexity didn’t make a big entrance. It wasn’t loud like ChatGPT, nor did it claim to “revolutionize everything.”
But in just two years, it’s quietly built a strong foundation:
Early 2023: ~1.5 million monthly users
End of 2023: 40 million monthly users
Mid 2024: Crossed 50 million monthly users
June 2025: ~22 million active users and over 160 million monthly page views, with 780 million+ queries per month
Numbers like these would make any startup proud. And yet—many still ask: “Wait, what’s Perplexity?”
At its core, Perplexity rethinks what AI can do for search—not just chat.
While ChatGPT was designed as a generative tool—writing essays, poems, and code
—Perplexity has always positioned itself as an information assistant. Its mission?
“Redefine search.”
Unlike Google, which sends you to a pile of blue links, Perplexity gives you direct answers, with sources attached.
Ask, “What are the trends in the 2024 European auto market?”, and you don’t get links
—you get a full summary with citations, right there on the page.
And here's the kicker: Perplexity doesn’t rely on a single model. It uses multiple LLMs behind the scenes
—GPT-4, Claude 3, Mistral, and its own PPLX model—dynamically switching based on the query.
Feature | ChatGPT | Perplexity |
---|---|---|
Purpose | Creativity, writing, coding | Search, summarization, citation |
Strength | Flexible generation, memory | Reliable, cited, concise answers |
Weakness | Hallucinations, unclear sources | Lacks creativity, limited dialogue flow |
Model | GPT-4o | GPT-4, Claude, PPLX (dynamic) |
In short:
ChatGPT is your creator.
Perplexity is your research assistant.
That used to be enough of a difference.
GPT caught up. Fast.
In the past year, OpenAI aggressively expanded ChatGPT’s capabilities:
- Web browsing in real-time (with source links)
- Bing integration for reliable citations
- Faster response times (especially with GPT-4o)
- Search-optimized prompting and RAG pipelines
Suddenly, ChatGPT could do most of what Perplexity was built for—and do it with a smoother, more conversational UX.
So while Perplexity technically remains distinct, it no longer feels different to many users.
And in product strategy, perception often trumps performance.
It’s not because it’s a bad product. Far from it.
- The summaries are sharp
- The citations are clear
- The speed is excellent
- The model-switching is smart
- But here’s the hard truth:
"If everything works in ChatGPT, why use anything else?"
Perplexity is battling the most powerful gravitational pull in tech: platform consolidation.
Just as Chrome swallowed up browsers and Spotify devoured music apps, ChatGPT is becoming the “one tool for everything” in AI.
Functionality alone isn’t enough anymore. To thrive, Perplexity needs:
A Stronger Positioning
Become the go-to tool for research, academia, or analytical deep-dives.
Right now, it’s just "pretty good at search." That’s not enough.
A Better Interaction Loop
Instead of a static answer box, offer tools for source tracing, comparison, and logical navigation through knowledge—something ChatGPT doesn’t do well yet.
Expand B2B Use
There’s serious value in positioning Perplexity as a tool for:
- Market research
- Competitive intelligence
- Policy analysis
- Academic literature review
Plug Into Ecosystems
Integration with Notion, Obsidian, Slack, or Microsoft 365 could create workflows that ChatGPT isn’t optimized for.
If Perplexity becomes part of how people work, not just how they search, it wins.
Perplexity is objectively solid. It delivers fast, trustworthy, well-sourced answers. It feels focused.
But it's playing a game where ChatGPT sets the rules—and keeps changing them.
If Perplexity wants to survive this AI arms race, it must move beyond “being better at search” and
embrace a unique identity, user journey, and purpose.
Because in 2025, “just working” isn’t enough. You have to matter.