The Age of the Answer Engine
Why traditional search is dead and what it means for knowledge work.
The integration of Gen AI into search engines is still in its formative phase. Major players like Google and Bing have worked to build models into their search interfaces with mixed success and overall, very poor user experience. The reality is these search engines' underlying business models and reliance on SEO-based bidding from advertisers make them unlikely to build a truly revolutionary search experience. That means the search industry is ripe for disruption.
Enter Perplexity. Launched in 2022, this AI-powered answer engine has quickly become a major disruptor to the search industry. It has already captured 10 million monthly active users and a $1 billion valuation. Why? Because it's tapping into a universal frustration: the mind-numbing inefficiency of current online research methods.
In this article, we will explore use cases for Perplexity, highlighting how and why it is such a transformative tool for knowledge workers. We’ll also look at what disruptive technologies like Perplexity mean for the future of knowledge work more broadly. More specifically, we'll examine:
The difference between traditional search vs Perplexity search
The impact on knowledge work and organizations
The broader implications for information access and synthesis
New skills that may become valuable as this technology evolves
Challenges that might arise as we rely more on AI-curated information
Part 1: Traditional Search vs Perplexity Search
By comparing traditional search methods with Perplexity's approach, we'll see firsthand why this tool is gaining traction and what it could mean for the future of knowledge work.
To illustrate the point, let’s look at the traditional search process and compare it to Perplexity. We’ll break it down with a real-world scenario.
The situation: You're a marketing strategist tasked with understanding your new client, Sunglass Hut. You need to get up to speed on the company and category quickly, so you decide to conduct some desktop research.
If you’re like most humans living on this planet, you would normally do what most of us have done over the past two decades, you turn to Google for answers. That research process, as you know, looks something like this:
Google "Sunglass Hut company history and background: Scroll past ads to find a mix of company website info and third-party articles. Open multiple tabs to compare information sources.
Look up "Sunglass Hut financial performance": Encounter paywalls for detailed financial reports. Find fragmented data from various financial news sites.
Search "sunglass retail market size and trends". Wade through sponsored content from industry publications. Open several tabs with conflicting market size estimates.
Google "top competitors of Sunglass Hut". Find listicles with varying opinions on who the main competitors are. Start separate searches for each potential competitor.
Investigate "customer preferences in sunglasses 2024". Scroll through numerous blogs with subjective trend predictions. Struggle to find credible, data-backed consumer insights.
Two hours later, you find yourself with 20+ open tabs and little to no clear picture of the category.
It’s as painful to conduct this process as it is to read it. But this is the tool we were given, until now. Perplexity flips this script. Instead of fragmented searches, you're engaging in a focused dialogue with an AI that's actually trying to solve your problem.
Here’s what a research flow looks like when done with Perplexity:
Since you know the client and the task, instead of searching, you go over to the Collections tab. Collections allow you to organize and group related Threads into folders and power your search with a pre-build set of instructions. Here’s what a good prompt for training a Collections tool looks like.
After clicking Create, we can go to the Library section and see that a Sunglass Hut Research Collection has been created.
From here, we click on the Sunglass Hut Research Collection, opening the project dashboard.
Here is where you start a new Thread, which allows you to group search topics into a long-form conversational search interface that blends synthesized Gen AI outputs and a host of rich media and links for the research.
Remember, it’s an answer engine, not a search engine, so queries can be much more complex and structured than traditional search. For prompt structuring, understand that Perplexity executes tasks similarly to the way an autonomous agent would, where the tool will identify the tasks outlined, build queries for each of them separately, mine multiple internet sources, and then generate a comprehensive and structured response with embedded resources, media, and links.
You can see in the search below how my prompt is structured. I break out a series of questions relating to the company’s background. This allows for precise and comprehensive outputs:
Of course, this is still a high-level overview. But this is where users can continue the conversation, diving deeper by asking more probing questions and reflecting on the material being surfaced.
Now let’s look at what the dashboard looks like after I’ve conducted a decent amount of research. (Tip: I used Claude 3.5 Sonnet to help me generate a structured and exhaustive set of 10 core areas to research, requested they be structured into independent chain-of-thought prompts, and then dropped them into individual threads to organize each core area). You can see some of these core areas in the Threads column at the bottom.
The 10 core areas were outlined and bucketed into a separate Thread within ~8 minutes.
The search quality is incredible. The depth of research and ability to continue the conversation in threads makes it easy, intuitive, fast, and focused. But now you have all of this amazing research, and you need to share it out. Remember when you would take notes, copy links and blocks of text, take screen grabs, jam it all into a Word doc, and send it out to the team? Not with this tool. With Perplexity, users can share collections with team members, and they can easily hop into the dashboard and read the research conducted.
But what if you wanted this to be even more interactive? It’s not very fun for you or your team to comb through a series of threads to read the research. That’s what Perplexity Pages are for.
They essentially look and function like a Wikipedia page. Users can curate them incredibly fast and convert their findings into interactive and rich media.
To do this, click on a Thread in your collections window, navigate to the top-right set of widgets, and select “Convert to Page”. You can easily add tabs, modules, videos, and images with very simple prompts. I like to think of this as an interactive, collaborative, immersive, and highly curated brief for teams.
As a disclaimer, Pages a bit clunky right now. Users can only create a Page for one thread at a time, meaning all of the Threads put together under specific Collections would need to be built in separately. Not ideal, but Pages are very new, so this should change soon. The workaround for this is to save your prompts and copy them into the page builder tab.
The main takeaway is that this type of workflow is more collaborative, intuitive, and immersive. It’s also much more rewarding. As Perplexity and other search tools continue to build on these technologies, we now have a good understanding of the direction in which it’s heading, and what that means for workers and organizations.
Part 2: The impact on knowledge work
Tools like Perplexity are not just changing how we find information; this isn’t “upgraded search” as I hope the example above illustrates. These tools are helping redefine our access to and synthesis of information. By augmenting the process of information gathering and initial synthesis, it's pushing knowledge workers up the value chain. Instead of spending hours collecting and organizing data, professionals can now get to analysis and solution formulation faster and more effectively.
Collaborative work will evolve as teams use shared dashboards as a common knowledge base. Similar features such as Claude 3.5 Sonnet’s Artifacts show this potential as well.
Key Implication: Knowledge workers will need to develop their metacognitive skills – understanding how to frame questions, interpret AI-generated insights, and apply them creatively to real-world problems.
Thinking about the second- and third-order effects, tools like Perplexity could potentially reshape our understanding and access to expertise. By providing coherent, comprehensive answers to complex queries, these tools lower the barrier to entry for specialized knowledge. This democratization of information could lead to more informed public discourse on complex issues and encourage broader, interdisciplinary thinking.
However, this ease of access to information also presents new challenges. As AI systems curate and synthesize vast amounts of data, there's a risk of perpetuating existing biases or creating new ones. The sheer volume and apparent authority of AI-generated answers might lead some to accept information without proper scrutiny.
Moreover, while these tools make it easier to access information, they don't automatically confer the ability to understand context, recognize nuances, or apply knowledge effectively. There's a risk of creating a false sense of expertise, where individuals have access to information but lack the depth of understanding that comes from years of study and experience.
Key Implication: The ease of access to high-quality information necessitates a greater emphasis on digital literacy and critical thinking skills. Users must learn to navigate potential biases in AI-curated information, question sources, understand context, and critically evaluate the information they receive.
In the age of answer engines, the skillset of knowledge workers is evolving rapidly. The focus is shifting from simple information gathering to sophisticated information evaluation and creative application. This trend, already visible with AI tools like Claude and ChatGPT, is amplified by platforms like Perplexity.
Critical thinking takes center stage as professionals must now assess the relevance, reliability, and context of AI-generated insights. The ability to formulate nuanced, probing questions becomes crucial – the depth and quality of answers directly correlate with the sophistication of queries. As these tools make cross-disciplinary connections more accessible, interdisciplinary thinking will become increasingly valuable, allowing professionals to uncover novel solutions and insights.
However, the true differentiator will be the human ability to go beyond what the AI provides. It's not just about understanding the information, but about creatively applying it to real-world problems, often in ways the AI might not anticipate.
Key Implication: High performers will be distinguished by their creativity in applying AI-generated insights to solve complex problems. Leadership will increasingly be defined by the ability to synthesize these insights with human intuition, experience, and strategic thinking, bridging the gap between AI-provided information and real-world application.
In conclusion, tools like Perplexity are ushering in a new era of knowledge work, transforming how we access, synthesize, and apply information. As these technologies evolve, they promise to democratize expertise while simultaneously demanding higher-order thinking skills from users. The future of knowledge work will belong to those who can masterfully navigate this AI-augmented landscape, combining the power of advanced answer engines with human creativity, critical thinking, and strategic insight.