Perplexity launches Computer for finance pros with fully traceable data

If you hand a portfolio manager a pitch with a hallucinated number, your career at that shop is effectively over. Accuracy isn't a feature in finance; it is the entire product.

Perplexity just launched Computer for Professional Finance, a dedicated platform tailored specifically for the research, analysis, and execution loops finance teams run every single day.

Bring your own terminals, or tap into 14 built-in feeds

The tool tackles the most annoying part of modern financial analysis: data fragmentation. Instead of making users rely solely on web scraping, it uses MCP connectors so teams can plug in their existing paid licenses from Morningstar, PitchBook, Daloopa, and Carbon Arc.

For shops without six-figure terminal budgets, Computer packs built-in finance tools pulling straight from 14 data providers, including Quartr and Fiscal. A hedge fund can run a complex portfolio breakdown using its proprietary feeds while simultaneously pulling live pricing from built-in trackers. The system also remembers a team's source preferences over time and automatically routes subsequent research tasks through those channels.

Perplexity also rolled out 35 pre-built workflows spanning ten specific sectors like Real Estate, Private Equity, and Public Equities. Instead of prompting an LLM from scratch, an asset manager can generate a cited one-page tearsheet into a PDF or slide deck, while a PE associate can build a live sourcing dashboard that teammates can remix without breaking the original prompt.

Every single metric links directly to the filing page

The real killer feature here is the audit trail. In finance, an unverified calculation is completely useless.

Inside Computer, hovering over any metric reveals its origin and timestamp. When a figure comes from an SEC filing and feeds into a downstream valuation calculation, the platform shows the exact math and links directly to the specific filing page. Crucially, this citation layer stays attached when you export charts, memos, or slides, saving junior analysts from spending hours rebuilding footnote trails before partner meetings.

The feature is available today inside Microsoft Teams and through the Agent API, with a native Excel add-in coming soon. On the FinSearchComp T1 benchmark for live financial lookups, Perplexity says its Finance Search Tool ranked first in accuracy, latency, and cost per correct answer.

Why it matters

For investment analysts and associates, this is arguably the most practical enterprise implementation of AI we have seen this year. It does not try to replace financial judgment; instead, it automates the brutal manual grind of cross-referencing SEC footnotes and assembling routine pitch books.

Strategically, Perplexity is playing this very smartly. Rather than trying to kill off Bloomberg or PitchBook, it uses open protocol connectors to sit right on top of them as the operating layer. By turning raw terminal data into finished presentations with audit trails, it attacks the software interface rather than the proprietary data moat.

The main question mark is reliability in edge cases. Financial modeling breaks quickly when non-GAAP adjustments or restated earnings get mixed up. Until the upcoming Excel plugin proves it can handle messy enterprise workbooks without subtle rounding errors, top-tier shops will treat these artifacts as impressive first drafts rather than unvetted ground truth.