Since its launch in December 2022, Perplexity has become a mainstay among AI power users – according to our most recent AI Proficiency Survey, 5% of the workforce uses Perplexity, but 33% of AI Experts do so. The company is forecasting $35 million in revenue this year, up from $5 million last year.

But it’s still relatively niche when compared to ChatGPT (61% adoption in our recent survey), Microsoft Copilot (32% adoption) and even Google’s Gemini (28% adoption). So this week, I talked to Perplexity power users and dug in myself to understand what power users are getting out of the platform.

What is Perplexity AI?

Think of Perplexity as an AI-powered search engine. Unlike Google, Bing, or the late Ask Jeeves (RIP), which provide a list of linked content in response to a query, Perplexity scours the web for you, and then uses AI to share answers in a more conversational way, while still citing its sources and including easy to QA links in its answers.

Here's a quick rundown of its differentiators:

• Conversational deep dives: Unlike traditional search engines, Perplexity allows you to ask follow-up questions, turning your search into a dialogue to help you get more personalized results.
• Assistant for web research: Think of Perplexity as an AI research assistant, summarizing key points from multiple sources and presenting a cohesive answer, saving you from tab-hopping.
• Citation trail: Unlike black-box AI responses via ChatGPT or Copilot, Perplexity leaves breadcrumbs. Every statement is followed by clickable source links, letting you easily see source material.

Pricing and features breakdown

Perplexity offers three versions of its product: Free, Perplexity Pro (for consumers), and Enterprise Pro (for businesses and enterprises). Here’s a breakdown of the pricing and features for each plan:

Quick vs Pro Search

Besides the ability to use multiple models, the biggest difference between Perplexity’s free and paid plan is the number of “Pro” searches available.

Quick Search is Perplexity's free option, designed for speed. It uses a less advanced LLM to provide shorter, more concise responses – think of it as a more intelligent Google search.
However, Pro Search is where Perplexity really shines. This advanced search offers more comprehensive results using powerful models from OpenAI and Anthropic. With Pro Search, you get:

1. Deeper research: Perplexity digs deeper into web information by aggregating more sources.
2. Better answers: Advanced models like GPT-4 and Claude 3 handle complex queries with more nuanced responses.
3. Multi-step reasoning: Complex questions are broken down into manageable steps, and Perplexity’s UI shows the “thought process” behind the answer.
4. Code interpretation and data analysis: Pro Search can run code directly in the chat interface for data analysis.

Beyond its emphasis on smart search, Perplexity offers two additional features that help you get the most out of its capabilities. Both are available in the free version of the plan.

The ability to “focus” your search

Perplexity offers focused searches in five topic areas for both quick and pro searches. This feature tailors results, providing personalized responses from relevant sources. You can choose from:

• Academic Focus: Searches academic databases and prioritizes scholarly, peer-reviewed articles in its search; great for students and researchers
• Social Focus: Scans social media platforms, forums, and discussion boards; great for marketers, researchers, or anyone looking to understand the public perception of a topic
• Video Focus: Searches YouTube and other video hosting sites to surface videos as well as video summaries and timestamps (all viewable in Perplexity); great for visual learners, recipe seekers, or researchers looking to analyze video content
• Writing Focus: This focus doesn’t search a specific area of the web, but is designed to help with drafting, brainstorming, writing, and coding; great for writers, content creators, and programmers
• Math Focus: Uses Wolfram Alpha to solve equations and conduct data analysis; great for students, researchers, and analysts

Perplexity’s replacement for Wikipedia

In May 2024, Perplexity introduced Pages, a feature designed to aggregate multiple sources into a readable article on a given topic.  It’s been dubbed the “AI Wikipedia” Here’s how it works:

1. You enter a topic and Perplexity searches the web for relevant information from sources.
2. Perplexity compiles the information into an article, automatically creating sections and subsections
3. Relevant images, tables, and other visual elements are incorporated to explain key concepts.
4. Every piece of information in the Page includes a clickable citation.
5. Related topics and follow-up questions are suggested at the end.

Users can edit the generated article, publish it for public viewing (similar to a Wikipedia page), or share the Page privately. Published Pages are living documents that can be updated at any time (again similar to Wikipedia). Take a peek at this page about who snagged the 2024 Nobel Prize in Medicine.

How we’re using Perplexity at Section

Unlike Claude and ChatGPT (both of which are used by 80%+ of our team), only 46% of Section employees use Perplexity. Therefore, it’s probably the most underutilized AI tool in our toolkit.

However, among the team using Perplexity, one use case shines through: Perplexity as Research Assistant. We promise students that all AI courses are updated every 30 days. Perplexity has become a huge part of this workflow.

We use Perplexity to find current and relevant data, headlines, or quotes to support our presentations and update our existing case studies. This leverages Perplexity’s real-time web search and citation sourcing, making it easy for us to fact check AI’s output.

Tasks that we’d typically give to an entry level research analyst, we now give to Perplexity. And our research analysts also use Perplexity themselves to save time. We’d estimate that for those using Perplexity, it’s replaced 10-20% of our Google search volume.

There’s too much information on the internet. Google organizes that information, but requires a specific boolean search to get the right source material, and still requires users to comb through multiple sources and aggregate insights themselves. Perplexity removes this manual labor and makes search more of a conversation.  

For individuals and teams that do a lot of research (aka Google power users), Perplexity is a game changer and might warrant the $20/month to increase access to deeper Pro searches and the benefits of multiple AI models. At a minimum, anyone using Google (aka everyone) should try Perplexity’s free version.