Vivold Consulting

Perplexity's 'Computer' pushes agentic AI into Microsoft 365 to pilot the desktop

Key Insights

Perplexity is expanding beyond search with Computer, an agentic orchestration system that pilots desktop files and local apps, now embedded across Microsoft 365 - Teams, plus add-ins for Word, Excel, PowerPoint, and Outlook. Built on its Comet browser, it routes tasks to multiple AI models inside a secure sandbox and reportedly did US$4.4bn of labor-equivalent work in three months. A macOS app, built with Intel, splits work between local hardware and the cloud.

Stay Updated

Get the latest insights delivered to your inbox

From a search box to a desktop conductor

Perplexity is trying to break AI out of the chat window with Computer, an orchestration layer that coordinates tasks across a user's apps and files rather than making them copy-paste between tools. It's the next step up from Comet, the company's Chromium-based agentic browser, and the headline move is how deeply it now reaches into Microsoft 365.

The Microsoft 365 push

Perplexity rolled the system into Microsoft's productivity suite in two stages: first inside Teams (early May), then as native add-in side panels in Word, Excel, PowerPoint, and Outlook (late May). The point is to let paying Pro, Max, and Enterprise users automate work without leaving the active window:

- In Excel, it can build a three-statement model from a 10-K pulled through a SharePoint connector.
- In Word, it can check a contract redline against a corporate template and generate an issues list.
- It also drafts PowerPoint decks from Excel charts and scans Outlook threads to assemble meeting briefs.

This matters because, before the integration, nearly 40% of Computer users were manually pasting AI outputs into Microsoft formats - friction the add-ins remove.

What's under the hood

Computer acts as a central coordinator that routes complex tasks to multiple leading models at once, running inside a secure sandbox with a real filesystem. Perplexity says it performed more than US$4.4bn in labor-equivalent work within three months of launch. A redesigned macOS app brings these agentic capabilities to local hardware through a hybrid framework built with Intel - routine tasks like text formatting run on the device's own processor, while heavier reasoning goes to the cloud, keeping sensitive files local.

The bigger ambition

The enterprise plumbing is there - SAML SSO, audit logs, granular admin controls, plus human-in-the-loop approval and source citation for every figure. The framing is unmistakably aggressive: for decades operating systems have been passive tool benches where people shuttle data between apps, and Perplexity is positioning Computer as a single AI coordinator that drives the machine on the user's behalf. Whether or not it truly displaces the desktop OS, it's a clear bet that the agent, not the app, becomes the primary interface.

Related Articles

An AWS knowledge-graph deployment turned 6-month research cycles into 3 weeks - and the blueprint transfers far beyond pharma

An AWS GraphRAG deployment in pharmaceutical research cut R&D cycles by 87% - initial discovery that took six months now closes in three weeks - by fusing siloed internal databases and public literature into one queryable knowledge graph on Amazon Neptune Analytics and Bedrock (running Claude). Every answer comes with verifiable citations and a mapped reasoning path, which is exactly what regulated industries need for compliance. The architecture is modular and, crucially, transferable: any enterprise drowning in fragmented legacy data can copy this pattern.

SpaceX, Anthropic, and OpenAI listings will out-value every US VC-backed exit since 2000 - reshaping vendor economics for everyone

The new NVCA-Pitchbook Venture Monitor dropped a stunning claim: the pending OpenAI and Anthropic IPOs, together with SpaceX's listing, will generate more value than every US VC-backed exit since 2000 combined. SpaceX is already public at $1.77 trillion, and with both AI labs pushing toward trillion-dollar debuts, the trio should land north of $4 trillion - against roughly $70 billion in total US IPO proceeds last year. For anyone buying AI services, the labs' shift to public-market scrutiny will reshape pricing, transparency, and vendor stability.

A 14-person open-source team just became the default way 8.9M developers run local AI - and a lever for slashing inference bills

Ollama, the open-source tool that lets developers run open-weight AI models on their own machines in minutes, raised a $65M Series B led by Theory Ventures ($88M total), revealing it now serves 8.9 million developers monthly and sits inside 85% of the Fortune 500 - with just 14 employees. Founders Jeff Morgan and Michael Chiang previously built Docker Desktop, and they're repeating the play: abstract away the hardware pain, then monetise a cloud tier priced on GPU time rather than tokens. The backdrop is the industry's loudest cost debate: every company with heavy inference bills is under existential pressure to shift routine workloads to open models.