Perplexity AI secures $73.6 million more for AI search

Perplexity AI Platform

Perplexity is a conversational AI platform that tunes search queries with follow-up questions, helping users get more relevant answers, according to the vendor.

However, compared to traditional AI chatbots, such as ChatGPT or Bard, and traditional Google search, Perplexity uses multiple LLMs, including Google Gemini, Mistral 7B, Anthropic's Claude 2.1 and OpenAI's GPT-4. It also uses its in-house model, Experimental.

Perplexity Pro subscribers pay $20 monthly to switch between the different models and can access various features, such as image generation.

Perplexity AI's fundraising success, which has the company valued at about $500 million, comes as many look for ways AI technology will change search.

In 2023, it appeared that Microsoft could shift the search market and uproot Google, especially when the tech giant incorporated ChatGPT into its Bing search engine.

Even Google, arguably the leader in search, has tried to revamp its search model with Experience, a generative AI-powered tool that helps users get answers to a topic or question.

However, the full potential of AI within search is far from being realized.

Transparency within search

With its AI-powered conversational search engine, Perplexity is addressing weaknesses of the current search model -- transparency and accuracy -- according to Nemertes Research CEO Johna Till Johnson.

"They've nailed an absolutely legitimate issue," she said.

They've nailed an absolutely legitimate issue.

Before the advent of AI technology, search engine users could type a question into Google or Bing and get multiple results. Users could then assess the validity of those results based on which result provided a primary source. With AI now incorporated into search, it's sometimes difficult to get valid results to questions because the AI technology could skew the results, Johnson added.

"What Perplexity is doing is saying, 'Well, we're going to give a front end that essentially addresses the question of where we got this information'" she continued.

The AI startup lets paid subscribers select not only the LLM they want to use but also where they want the source of their answer to emerge from. Users can choose to only have answers from peer-reviewed sources instead of just any source on the web.

"That's actually a big deal," Johnson said. "They're taking advantage of something that's really smart, which is now that Google and Microsoft have added AI to their search, they've turned it into junk."

Competing with Google

However, whether startups like Perplexity can shake up Google's free model and convince customers to use a paid search service is still to be seen.

"Google is a multibillion-dollar company that was built on search," Futurum Research analyst Mark Beccue said. "They have a significant franchise to protect."

While the AI technology behind Perplexity is strong and comparable to that of Microsoft Copilot, the only other company that has a shot at beating Google out of search is likely Microsoft, Beccue said.

"It's going to be hard for a smaller player to break in," he added.

It will be harder for a vendor that's asking users to pay for better results like Perplexity even though some users don't trust the current free search paradigm, he continued.

However, the startup could gain a following if it can deliver on source transparency and veracity, Johnson said.

For Perplexity, while the long-term goal is to disrupt Google's search model, the current goal is to scale and improve its platform, CEO Aravind Srinivas said in an interview.

"The immediate thing is to first perfect the experience," Srinivas said. "Produce less hallucinations, make sure the answers are as perfect as possible [and] make sure you can cover a wide range of search queries not just related to knowledge."

Once the startup optimizes its user experience, it then plans to monetize differently than Google. The startup doesn't want to depend on advertising revenue to monetize. It would like to find a different way, although that's still unclear, Srinivas said.

However, even as the market moves towards an ask-and-answer AI search experience, Google's search model -- with its ads and links -- will likely fade over time instead of being completely erased, Srinivas maintained.

"Everyone's going to realize this slowly over time as the accuracy of these AI bots increases and the cost of running these AI bots goes down and also the coverage of what the AI bots can handle [increases]," he added.

However, for now, such changes are still in the early stages, Beccue said.

"AI is going to change how we do search," he said. "It's just kind of formative. And we're really just at the front end of seeing what it will do."

Esther Ajao is a TechTarget Editorial news writer covering artificial intelligence software and systems.