Background agents
Background agents start on their own when you add a file. They sort and read what comes in, so Mona and its subagents have evidence to work from. You don't start them or talk to them, and you step in only when they ask.
| Job | What it does | Where you see it | When it needs you |
|---|---|---|---|
| Document sorting | Matches each PDF to a party, splits a bundle, and identifies each document's type | Documents | Bundle to split, Document type unclear |
| Document reading | Reads each document's dates, account, period, and contents, and keeps the page each came from | A document's details | Couldn't read |
| Transaction reading | Takes the balances and transactions from each account statement, checks that they add up, then categorizes each one and names its merchant | Accounts, Timeline, Missing accounts, and Cash flow | Transactions don't balance, Needs an account |
| Statement reading | Reads a Statement of Net Worth line by line, once you confirm its details | Statements | Confirm details, Reading stopped before it finished |
| Statement matching | Connects each line to the financial item it describes, either one already in the party's filings or a new one | Statements and Net worth analysis | Matching stopped twice |
Usage in your firm's settings lists this work by its own operation names, such as Document partitioning, Ledger extraction, and Statement extraction.
When they need you
- Documents: whatever a document needs is listed under Needs review, grouped by problem, with what to do about each.
- Statements of Net Worth: each statement's card shows its steps and what to do when one stops.
What they don't do
Reading tells you what a document says. Whether it supports a statement's figures is the work of Trace and the Statement agent.