Perplexity is rolling out a new version of its AI “Computer” agent, and this one is aimed squarely at the people who live in Excel, earnings calls, and data rooms all day: professional finance teams. Computer for Professional Finance is designed to sit on top of the data, tools, and workflows analysts already use and turn them into concrete work products – think tearsheets, memos, dashboards, and annotated charts – without losing the audit trail behind every number.

At a high level, Perplexity is trying to solve a problem that every analyst knows too well: you spend more time hunting for data, normalizing it, and re-running the same process for the hundredth time than you do actually thinking about the investment or decision. Computer for Professional Finance takes that grind and wraps it into reusable workflows that plug into your existing data stacks and collaboration tools.

Instead of acting like a generic chatbot, Computer behaves more like a research assistant that already knows where your data lives, which providers your firm pays for, and what kind of artifacts your team needs at the end of a process. Finance is already one of the top use cases for Perplexity, so this product is less about entering a new vertical and more about professionalizing a behavior that was already happening on the platform.

The foundation of this release is “trusted data” – which, in finance, is non‑negotiable. Perplexity lets firms plug in their existing subscriptions from providers like Morningstar, PitchBook, Daloopa, and Carbon Arc through managed connectors, so Computer can work directly on top of the same sources people already trust for portfolio construction, valuation work, or deal sourcing. If you’re a smaller shop without a Bloomberg or full data stack, Computer falls back to a set of 40-plus built‑in finance tools that pull from more than a dozen external providers, including platforms such as Quartr and Fiscal for fundamentals, transcripts, and market data.

That mix of “bring your own data” plus curated built‑ins makes the product flexible enough for different types of teams: a hedge fund can route everything through its licensed feeds, while a boutique PE firm or family office can still get a serious research stack without paying for a terminal. Over time, Computer remembers which sources each team prefers, so if your shop always defaults to a specific provider for credit data or ESG metrics, the agent learns to lean on those by default instead of forcing you to re‑specify preferences every time.

Where it gets more interesting is how all of this data turns into actual work. Computer for Professional Finance ships with roughly three dozen prebuilt workflows that match the way finance teams actually spend their weeks: deal screening, company deep dives, portfolio monitoring, risk analysis, and market mapping across segments like real estate, private equity, public equities, and asset management. A workflow is essentially a packaged playbook – it knows what data sources to use, how to structure the analysis, and which output format to generate, so instead of juggling prompts and spreadsheets, you tell it the objective and let it run the full loop.

Take a “Company Tearsheet” workflow as an example. An analyst can feed in a ticker or private company name plus a preferred data source, and Computer will compile a one-pager that covers financials, business overview, key risks, and recent catalysts, complete with citations, then render it as a PDF, slide deck, internal web app, or a shareable Computer thread depending on what the team needs. Because the whole thing runs on top of connected data sources, updating the tearsheet doesn’t mean redoing the work; you can simply re-run the workflow with updated dates or scenarios and get an up-to-date version using the same logic.

On the private equity side, a “Sourcing Screen” workflow takes high‑level signals – say, a revenue band, region, margin profile, and a couple of industry themes – and turns them into a live dashboard of potential targets that colleagues can filter and remix without overwriting the original work. Instead of sending around static Excel lists and email threads, the team can treat that dashboard as a living artifact, updating it as new companies, signals, or hypotheses come in.

For hedge funds and public markets investors, Perplexity highlights workflows like an “Annotated Stock Chart” that blends price data with events and news. The result is an interactive view of a security’s history with annotations for catalysts like earnings surprises, ownership shifts, or rating changes, making it easier to reconstruct “what actually happened” around a move rather than jumping between charting tools, filings, and news archives.

Underneath all of these workflows is a pretty strict stance on traceability. In finance, any chart or memo is only as valuable as the evidence behind it, and Perplexity leans into that by letting you trace every numeric value back to the underlying source. Hover over a data point and you can see where it came from, when it was fetched, and, in the case of SEC filings, exactly which filing page and line items fed into any downstream calculations.

That traceability isn’t just a UI flourish. It travels with the output itself – so if you export a memo, a deck, or a dashboard, the links back to the underlying evidence stay attached, which makes internal reviews and IC processes far less painful. Instead of analysts scrambling to recreate how they got from 10-K tables to a final model, reviewers can click back into the chain and see the math and sources in context.

Perplexity is also positioning Computer as something that “works where you work” rather than yet another standalone portal. Today, the finance‑focused version runs in Microsoft Teams and through the Perplexity Agent API, with an Excel add-in on the way. In Teams, people can pile into the same task thread, add context, and then review Computer’s output together, which is much closer to the way deal teams, PMs, and risk committees already collaborate than a single‑user chat interface.

For developers and quant‑leaning teams, the Agent API exposes a Finance Search Tool that handles time‑sensitive queries like live stock, crypto, and FX prices as part of an automated workflow. That tool has been benchmarked on FinSearchComp T1 – a recognized open benchmark focused on real‑time financial search – where Perplexity’s system ranks at the top on a mix of accuracy, latency, and cost per correct answer, which matters if you’re embedding it into systematic workflows that don’t tolerate stale prices or slow responses.

The Excel integration is the most obvious bridge between AI and the reality of finance work. Rather than forcing analysts to upload models or recreate logic in a new interface, the Excel add-in is built to read the existing workbook context, pull in connected licensed or built-in data, and generate research outputs directly from within the sheet. In practice, that could look like asking Computer to scan a multi-tab model for key drivers, build a risk section for a memo, or generate charts and commentary for a quarterly board pack, all without leaving the file that the team is already aligned on.

Zooming out, this release fits into a broader push by Perplexity to become a kind of “operating system” for financial decisions. The company has already positioned Computer as a personal CFO tool for individuals, letting users link their bank, investment, and loan accounts through Plaid and then analyze spending, track net worth, or model retirement scenarios in one place. On the institutional side, Computer for Enterprise is marketed as a multi-model, multi-step agent that supports legal, marketing, and other teams, with finance as one of the most demanding early adopters.

Computer for Professional Finance is basically the professional counterpart to those moves: it brings the same agentic engine, but wrapped in the governance, data connectivity, and auditability that institutional players expect. Large investment firms, retail brokerages, and major tech names already use Perplexity for financial work in some form; this product gives them something more structured to standardize on, instead of treating AI as an optional sidekick running in a browser tab.

Perplexity is also entering a competitive landscape where other AI companies are racing to automate more of the finance stack, but the company is betting that its focus on live search, source-level transparency, and model orchestration is a differentiator. In practical terms, that means Computer can pull the freshest data from the open web, route tasks to different underlying models depending on the job, and then surface a paper trail that looks more like a clean audit log than a black-box answer.

For finance professionals, the pitch is pretty simple: keep using the systems you already rely on – your data vendors, Teams, Excel, internal knowledge bases – and let Computer handle the repetitive research, synthesis, and formatting around them. You still own the judgment calls and the investment thesis, but you offload the drudgery of stitching together sources, running the same analyses by hand, and proving to colleagues where every number came from.

Computer for Professional Finance is available starting today for enterprise customers, with Perplexity steering interested teams toward its finance use-case pages and enterprise sales channels to get set up. The real test will be whether analysts and PMs feel comfortable trusting an AI layer as part of their everyday stack, but Perplexity is betting that combining licensed data, strong traceability, and tight integrations is enough to make that leap feel less like a risk and more like an obvious upgrade.