Perplexity CEO Unveils OpenClaw-like AI Tool for Non-Experts

Perplexity’s New Direction

Recently, AI enthusiasts have noticed a shift: Perplexity CEO Aravind Srinivas, once a vocal presence on social media, had gone unexpectedly silent. His quietness stood out in a time when agent-style tools like Claude Code, Codex, and the popular open-source OpenClaw were gaining attention. Perplexity, known as an AI-powered “answer engine” and competitor to Google Search, seemed to be on the sidelines. Some even speculated that the company might have lost its way.

However, according to Srinivas, Perplexity was not astray but rather deeply focused on development. I had a conversation with him shortly after the launch of Computer, which aims to transform the current powerful yet complex agent tools into a more user-friendly digital workspace suitable for non-experts. Initially, this product is available exclusively for Perplexity Max subscribers, with plans for a wider release for Pro and Enterprise users in the upcoming weeks.

An Accessible Solution

Computer functions like an OpenClaw for everyone, allowing users to handle tasks such as research, writing, or coding without giving full control of any personal device to an AI. The defining characteristic of Computer is its model-agnostic approach: various components of a task are assigned to the most effective model. Presently, it coordinates an impressive array of 19 models on the backend, including Claude Opus 4.6 for orchestration and coding, Google Gemini for in-depth research, and ChatGPT 5.2 for extensive context retrieval.

“When assembling a team, you don’t create a uniform group with identical skills,” Srinivas noted. “Instead, you gather individuals with diverse strengths. We’re applying that concept to AI workflows; the orchestration is the core product, while the models function as tools.”

This multi-model strategy is not entirely new for Perplexity; more than half of the company’s enterprise users already choose from multiple models within a single day. With Computer, this philosophy has become a central tenet and a strategic advantage. Srinivas remains unfazed by the potential restrictions from model providers like OpenAI or Anthropic: “I even received congratulatory messages from Anthropic and Google,” he mentioned, emphasizing the mutual benefits that arise when their systems contribute to broader workflows. If access parameters evolve, Perplexity is ready to pivot. “The model landscape is more competitive than ever,” he remarked.

Safety and Usability

Srinivas contrasted Perplexity’s approach with tools like OpenClaw, which often run locally with extensive access to files and settings—an approach that can resemble malware due to its potential to compromise data security. Perplexity’s system operates in the cloud within a secure environment, performing tasks in the background. This method resembles assigning work to a colleague on Slack, thereby cultivating a safer and more reliable system.

While Srinivas mentioned more ambitious features are on the horizon for Computer, the primary focus remains on accessibility. He expressed that with Computer, “Even your mom could text on the app to delegate tasks,” while using OpenClaw requires careful configuration by engineers. Perplexity seeks to make agent-style work feel intuitive, akin to operating a Macintosh or an iPhone, rather than wrestling with server setups. “It genuinely feels like I have a swarm of agents working for me,” he said, describing an experience that felt more profound than mere buzzwords.

AI Highlights

Here’s the latest in AI news.

Sharon Goldman
sharon.goldman@fortune.com
@sharongoldman

FORTUNE ON AI

AI capital expenditure and the ‘wealth effect’ from technology stocks (like Nvidia) now account for one-third of U.S. GDP growth, according to top analysts – by Jim Edwards

What AI bubble? Nvidia reports a record $68 billion quarterly revenue and a $78 billion forecast, with Jensen Huang citing skyrocketing agent adoption – by Amanda Gerut

Tech companies are projected to invest an unprecedented $700 billion in AI data centers this year; Nvidia’s Jensen Huang warns we are far from hitting our peak – by Alexei Oreskovic

In its struggle with the Pentagon, Anthropic confronts a major crisis in its five-year history – by Jeremy Kahn

Exclusive: Bilt’s new AI ‘Neighborhood Concierge’ challenges Amazon and other e-commerce giants – by Diane Brady

Exclusive: A startup targeting Nvidia’s dominance in AI data center workloads secures $10.25 million in venture funding – by Jeremy Kahn

AI IN THE NEWS

AI politics heats up as Florida Governor Ron DeSantis embraces skepticism to differentiate himself from Vance. Over the past year, I’ve been exploring the political ramifications surrounding AI, and it’s fascinating to observe the rift between Florida Governor Ron DeSantis and his Republican counterparts—particularly Vice President JD Vance. According to NBC News, he is taking a unique stance in U.S. politics by being skeptical of AI, cautioning that the rapid growth of AI—particularly power-hungry data centers that receive public funding—could escalate energy costs, jeopardize jobs, and harm the environment. Presenting himself as a protector of consumers, DeSantis warns that large data centers can consume as much electricity as a medium-sized city, insisting that Floridians shouldn’t be left “with the bill.” His viewpoint aligns with public sentiment as recent polls suggest that 63% of Americans believe AI will lead to job losses, while 41% of voters support banning data centers in residential areas, signifying a growing unease over AI’s impact, despite leadership and industry calls for advancement.

Mistral AI secures Accenture as a significant partner. Mistral AI, a European model developer that has been somewhat overlooked in recent discussions, achieved a major milestone by partnering with Accenture as its newest ‘systems integrator.’ According to the Wall Street Journal, under a multi-year agreement, Accenture will utilize Mistral’s models and solutions for enterprise customers as they transition from AI pilots to full-scale implementation. This collaboration adds to a list of clients that already includes notable names like IBM, Cisco, SAP, Stellantis, and ASML, which invested over $1.5 billion for about an 11% stake last year, valuing the startup at nearly $14 billion. This partnership underscores how consultancies are increasingly influencing which AI models become pivotal in real-world applications. Notably, Accenture is also collaborating with OpenAI to assist clients in integrating its Frontier enterprise AI agent platform and maintains a multi-year partnership with Anthropic.

AI exacerbates existing internet privacy challenges.An intriguing New York Times article highlights recent incidents that are intensifying discussions about privacy and accountability in the era of AI chatbots and assistants. A recent ruling revealed that conversations with Anthropic’s Claude do not enjoy attorney-client privilege, while a controversial Super Bowl ad from smart doorbell company Ring sparked criticism for its promotion of AI-enabled neighborhood surveillance. Additionally, OpenAI disclosed it had previously reviewed a user’s ChatGPT messages prior to a tragic shooting, prompting fresh debates on the circumstances under which AI companies should disclose private conversations to authorities. Experts argue that while the underlying risks surrounding data access aren’t new—data shared with companies has long been accessible in certain contexts—chatbots create new stakes by encouraging individuals to disclose more intimate information than traditional tools. This change is likely to intensify as AI agents gain access to users’ emails, messages, and devices.

EYE ON AI NUMBERS

57%

This figure represents the percentage of U.S. teenagers who have utilized AI chatbots for information-seeking, according to a recent Pew Research survey involving 1,458 teens aged 13–17, released this week. The study reveals that these tools have become integral in teen life, with 54% claiming to use chatbots for school assistance, 47% for entertainment, and 10% saying they rely on AI for most or all of their schoolwork. Concurrently, 59% report that AI-related cheating occurs in their schools at least occasionally. Despite this, opinions on the technology are generally optimistic, with 36% foreseeing a positive impact from AI in their lives over the next two decades, compared to 15% expecting a negative outcome.

AI CALENDAR

Feb. 24-26: International Association for Safe & Ethical AI (IASEAI), UNESCO, Paris, France.

March 2-5: Mobile World Congress, Barcelona, Spain.

March 16-19: Nvidia GTC, San Jose, Calif.

April 6-9: HumanX, San Francisco.