Most portfolio monitoring reports the past. Powdr reads the forecast.

You see the breach before it happens

Forward-looking covenant monitoring, not backward-looking certificates.

Powdr tests every covenant against every month of the forecast, not just the last reported quarter. A covenant that passes today but fails in four months’ time isn’t marked compliant — it’s marked watch, and Powdr names the month it breaks. Headroom is calculated as slack against the limit and knows which way each test cuts, so a leverage covenant and a minimum-cash covenant are both measured correctly.

Every number shows its working

Click any figure and Powdr tells you where it came from.

Each metric carries its own provenance: which model rows it reads, where to find them in the model, the arithmetic in plain business language, which month it’s struck at, and why a correct figure might differ from the same sum done by hand. The internal risk grade works the same way — it’s a transparent 1-to-10 scorecard that returns the list of reasons it notched a borrower up or down.

And when a model genuinely can’t support a figure, the tile says why instead of printing a zero.

Sponsor returns maths that survives the valuation committee

IRR, MOIC, TVPI and DPI on every stake — plus a full exit waterfall.

Powdr’s IRR is a real XIRR: actual/365 day counting matching Excel, solved by bisection and Newton-Raphson with an analytic derivative, because a finite-difference approach loses too many significant digits on flows of tens of millions. Throughout the returns and valuation engine, money is held in arbitrary-precision decimal arithmetic rather than floating point.

The waterfall models prioritised debt tranches, transaction costs, preference claims, hurdle IRRs, catch-up and participation, and reports each holder’s capital returned and money multiple — plus any preference shortfall.

The waterfall runs a company at a time. It is the fund-level aggregation of IRR, MOIC, TVPI and DPI that spans the whole portfolio.

One platform that speaks lender or sponsor

A switch in the header changes the metrics, the layouts and the vocabulary.

A lender and a sponsor are not asking the same question, so Powdr doesn’t pretend they are. In lending mode a collection of companies is a sub-portfolio and the whole thing is a book; in private equity mode the same collection is a fund and the whole thing is a portfolio. Each mode keeps its own dashboard layouts, so switching back and forth costs nothing.

Getting the data in doesn’t mean re-keying it

Upload it, or connect the accounting system — then let AI do the mapping and check its own work.

Send Powdr a spreadsheet, a PDF or even a PowerPoint pack and it reads the financial statements out of it, classifies every line into a driver category, and proposes the mapping for you to review. Or connect Xero, QuickBooks Online, Sage or FreeAgent directly — yourself, by reusing an existing connection, or by emailing a single-use invitation to the company’s bookkeeper and letting them authorise it without ever seeing your portfolio.

Then Powdr checks itself. It reconciles its own model against the source document line by line and month by month, explains each difference, and where a mapping change would close the gap, offers it as a one-click fix behind an automatic snapshot. It will even tell you when the discrepancy is a broken formula in the source file rather than an error in the model.

Accounting refreshes are triggered when you ask for them, which is why the data freshness view exists.