Perplexity launches Computer for Enterprise, claims $1.6M labor savings in internal test
Perplexity made Computer for Enterprise generally available to enterprise customers on March 12, claiming an internal study of 16,000+ queries showed $1.6 million in labor cost savings and 3.2 years of equivalent work completed in four weeks. The service integrates with Gmail, Outlook, GitHub, Linear, Slack, Notion, Snowflake, Databricks, and Salesforce, orchestrating tasks across 20 frontier models with agentic internet access.
Perplexity Launches Computer for Enterprise, Claims $1.6M Labor Savings in Internal Test
Perplexity released Computer for Enterprise on March 12, extending its AI agent platform to enterprise customers. The service integrates with nine major enterprise platforms and claims substantial productivity gains based on internal testing.
What Computer for Enterprise Does
Computer for Enterprise is a cloud-based orchestration layer that runs background tasks using AI models, conditional triggers, web research, and sub-agent delegation. It connects to Gmail, Outlook, GitHub, Linear, Slack, Notion, Snowflake, Databricks, and Salesforce, allowing automated interaction with data stored across these platforms.
The service combines "highly accurate AI search, an orchestration harness of 20 frontier models, and agentic internet access," according to Perplexity. Task examples include: triaging support tickets by severity and drafting customer responses; automating due diligence by fact-checking claims against live data and flagging inconsistencies; and generating annotated reports.
Claimed Productivity Gains
Perplexity claims an internal study involving 16,000+ queries showed Computer for Enterprise saved internal teams $1.6 million in labor costs and completed the equivalent of 3.2 years of work in four weeks. The company did not disclose methodological details, benchmark definitions, or how labor value was calculated, making independent verification impossible.
Personal Computer Launch
Simultaneously, Perplexity launched Personal Computer on March 12, a local complement to its cloud-based service. Personal Computer runs on a dedicated Mac mini, operating 24/7 and connecting to local applications while maintaining secure server connections. The company claims sensitive actions require approval and that full audit trails and a kill switch provide user control.
Perplexity's positioning—that "AI is the computer" in the cloud version, while your actual computer is the computer in the local version—appears designed to differentiate the two offerings and justify the "Computer" branding across both products.
Enterprise Adoption Hurdles
The launches occur as enterprises remain cautious about delegating tasks to software agents. Security concerns persist despite Perplexity's claims about approval workflows and audit trails. The company provided minimal technical detail on how sensitive data is handled, encryption mechanisms, or compliance frameworks.
Access to Personal Computer remains on a waitlist basis. Computer for Enterprise pricing was not disclosed.
What This Means
Perplexity is positioning itself as an enterprise automation platform competing with offerings like OpenAI's agents capabilities and enterprise RPA tools. The $1.6M savings claim is unsubstantiated without methodological transparency, but it signals Perplexity's confidence in the product's productivity impact. The real test will be whether enterprises trust AI agents with access to their Gmail, GitHub, Salesforce, and Snowflake instances—a significant security and governance question the company has not adequately addressed in public statements.