aFA Agent — AI-native fund accounting
Built for capital market asset classes with liquid pricing and trading and sound intermediaries. An append-only ledger produces your fund's shadow NAV from source events, ahead of the official number from your administrator. AI maps files, flags anomalies, explains variances, and drafts commentary — but never writes a journal or a NAV figure itself. A named human on your team reviews, decides, and signs off on every result.
Great for single NAV calculation, not series or equalization.
aFA Agent works alongside with your lean team, not a replacement for the fund administrator producing your official NAV. It gives you an internal, independent shadow NAV — so you know what your books say before the official number lands.
aFA Agent self-ingests files or connects via API with your intermediaries. Your operations staff review exceptions and anomalies, your fund accountant runs fees and the ledger, and your manager reviews and signs off — producing an internal estimate NAV before the official NAV arrives from your administrator.
These principles hold across every module below — they are what your own auditor will check first.
Nothing is ever overwritten. A correction is a new entry that supersedes the old; the old stays visible — which is what lets any past NAV be rebuilt to the cent.
The double-entry ledger is authoritative. The balance sheet and P&L are views of it, not separate records to reconcile.
Same inputs always produce the same NAV, and any past period re-runs identically. AI must never write a journal or a NAV figure.
AI maps files, flags anomalies, explains variances, and drafts commentary. A named human on your team reviews, decides, and intervenes on any exception.
Ownership, accountability and liability always rest with a person on your team. Every automated action is reversible and logged with a named approver.
The data model carries the structure to support more complex funds as your needs grow, without needing to be rebuilt.
From fund setup to a signed-off NAV — grouped by what each capability delivers, not how it's built.
Your fund's legal structure, share classes, terms and investor register in one place. AI flags gaps against your offering document and verifies subscription and redemption records before they hit the ledger.
Reference data for the securities and brokers or custodians you trade with — enriched and de-duplicated automatically, so your ledger always books against clean, current records.
Broker, custodian and price files are imported and mapped automatically, in whatever format they arrive. Anything AI can't confidently map is queued as an exception for your team, not silently guessed.
Trades become positions with cost basis and P&L; prices and FX rates are versioned and checked for staleness; every entry lands in a double-entry, append-only ledger. AI catches duplicates, off-market trades and unusual entries before they become a problem.
Management fees, expense allocations, amortisation schedules and income accruals — calculated against your fund's actual terms and rechecked automatically each period.
Your shadow NAV moves through estimate, final and confirmed status with anomaly checks at every stage. Once your administrator's official NAV lands, it's reconciled automatically against your shadow number, with AI explaining what drove any variance.
Balance sheet, P&L, holdings and cash-movement reports regenerate on demand, with AI-drafted commentary ready for your team's review before anything goes out.
Every entry, edit, AI suggestion and human approval is logged and immutable. Corrections and scenarios can be tested in a discardable copy of your fund's state before anything is posted live.
The 95–99% accuracy target isn't a marketing figure — it's measured against a permanent internal test suite covering the structures and mechanics real funds actually use.
Single NAV accounting handling across multiple share classes under a Cayman Limited fund structure.
Class-level P&L attribution and illiquid asset marks under a Singapore Limited Partnership structure.
Uneven profit allocation and derivatives accounting across investor classes with different profit-sharing terms.
Accuracy is measured as shadow NAV within 95–99% of the administrator's official NAV, produced same-day to T+2 — the standard your own reconciliation would apply.
Tell us about your fund structure and your current accounting process, and we'll walk you through fit.
Because the NAV is a number your own management relies on, aFA Agent's governance model is strict: AI is never the author of a financial figure.