How Perplexity has been built. In today’s newsletter, Perplexity's $18 Billion Answer Engine will be broken down.

New Cohort starting on 3rd September 2025: AI Engineer HQ

In today’s edition:

• AI Case Study— How Perplexity Built an Answer Engine to Challenge Google?

• Build Together— Here’s How I Can Help You

[Sponsor Spotlight]

Start learning AI in 2025

Keeping up with AI is hard – we get it!

That’s why over 1M professionals read Superhuman AI to stay ahead.

• Get daily AI news, tools, and tutorials

• Learn new AI skills you can use at work in 3 mins a day

• Become 10X more productive

Sign up and start mastering AI

[AI Case Study]

How Perplexity Built an $18 Billion Answer Engine to Challenge Google?

4 Chapters You and I are going to touch:

1. The Search Dilemma

2. How does Perplexity’s Engine Work?

3. The Business Model

4. 2 Big obstacles threaten Perplexity’s future

The Search Dilemma

For years, users faced a frustrating choice online:

• search engines gave endless blue links filled with ads, leaving users to piece together their answers

• AI chatbots provided confident responses, but often presented incorrect facts and rarely cited sources

Perplexity.ai, founded in 2022, took a different path.

They created an answer engine that blends the web’s real-time knowledge with the reasoning of LLMs.

The result:

• direct

• accurate

• cited answers

By mid-2025, this approach propelled the 100-person startup to:

• $18 billion valuation

• $100 million annual recurring revenue

• a reputation as Google’s most credible challenger

How does Perplexity’s Engine Work?

Perplexity’s success rests on three pillars:

1. RAG Engine

2. Model Router

3. From answer to intelligence

1) RAG Engine

Trust through transparency.

Their core innovation is simple - every answer comes with citations.

This builds trust and addresses the credibility issue associated with AI.

How it works:

• the system interprets the query and pulls fresh information from the web

• content is broken into smaller pieces, stored, and re-ranked by relevance

• the language model produces an answer and links to the sources

This process grounds answers in real data, ensuring accuracy and accountability.

2) Model Router

Pragmatism over purity.

Instead of betting on one model, Perplexity routes queries to the best option.

• simple queries - fast in-house Sonar-7B model

• complex reasoning - Claude 3.5 Sonnet via Amazon Bedrock

• coding/creative tasks - GPT-4o

This mix lets them optimize for speed, cost, and performance all at once.

crawler → chunker → embeddings → vector DB → retriever → prompt → LLM → citations

3) From Answers to Intelligence

Perplexity expanded beyond Q&A to build an ecosystem.

• perplexity pro ($20/month) access to premium models and higher limits

• deep research agent - creates structured research reports in minutes, attracting enterprise interest

• enterprise pro ($40/seat) collaboration tools, like Spaces for internal knowledge management (NVIDIA and Databricks)

This suite turned casual users into paying customers and businesses into long-term clients.

2 Big obstacles threaten Perplexity’s future

1 - Copyright lawsuits

• publishers like News Corp are suing over web crawling

• outcome may shape how AI can access online content

2 - Infrastructure costs

• handling 400 million queries monthly is expensive

• with costs at about $0.046 per query, they rely on optimization, caching, and cloud deals to stay sustainable.

Key Takeaways for Leaders

• bitations and transparency win credibility

• the best experience often comes from mixing models, not insisting on one solution

• start with a strong feature, then expand into a suite that deepens user engagement

• perplexity scaled with just 100 people by shipping fast, measuring results, and iterating constantly

Conclusion

Perplexity.ai proved that in the age of AI, trust and execution matter as much as technology.

With a lean team, smart strategy, and relentless focus on accuracy, they built a product that challenges one of the biggest tech monopolies in history, aka Google.

Their story shows that even in a field dominated by giants, a focused team with the right approach can redefine the market.

Until next time.

Happy AI

Before you go: Here’s How I Can Help You

• AI Engineering & Consulting (B2B) at Dextar—[Request a Brainstorm]

• You are a leader?—Join [The Elite]

• Become an AI Engineer in 2025—[AI Engineer HQ]

• AI Training for Enterprise Team—[MasterDexter]

• Get in front of 5000+ AI leaders & professionals—[Sponsor this Newsletter]

I use BeeHiiv to send this newsletter.

PS: Which case study do you want next?