← Main site LedgerIQ Private beta
Book a walkthrough
How It Works

From paperwork to posted ledger

The pipeline, the 20 inputs, and where the human sits.

LedgerIQ is one pipeline: every input format lands in the same place, becomes a double-entry draft, passes a human check when it matters, and posts to an immutable ledger that feeds the dashboards and exports.

1. Twenty ways in

Businesses do not produce clean data — they produce receipts, photos, PDFs, spreadsheets, statements and voicemails. LedgerIQ takes all of it. Bank and card feeds connect directly; the rest is captured by photo, upload or forward.

CategoryExamplesHow it comes in
Physical documentsPaper receipts, vendor invoices, handwritten notes, register tapes, petty-cash logsPhoto / scan → extraction
StatementsBank statements, credit-card statements, processor settlementsDirect feed (Plaid, Stripe) or upload
OperationsPOS daily sales, e-commerce orders, inventory POs, payroll reportsPOS / commerce / payroll connections
ConversationalExpense voice memos, in-app chat context, mileage logsRecord, type or forward
ComplianceW-2 / 1099 tax documents, owner-drawing notesUpload → extraction and flagging

Bank feeds (Plaid) and processor-settlement feeds (Stripe) run on real integration code and activate as soon as credentials are provisioned for an engagement. During the beta, some connectors are demonstrated with representative data — the walkthrough shows you exactly which are live for your setup.

2. AI turns it into a draft entry

3. A human reviews what matters

Not every entry needs a person, and not every entry should skip one. LedgerIQ routes by risk:

Entry looks likeWhat happens
High-confidence, known vendor, routine amountMoves through to posting
Low confidence, unknown vendor, or high dollar valueRoutes to the trained accountant back-office queue
AI cannot place a payment at allA plain-language clarification goes to the business owner; one answer clears it

The back-office reviewers are trained accountants doing ledger review — LedgerIQ provides the ledger and the tooling, not tax or advisory work. The state machine is explicit: queued → processing → review → posted, with posting as a one-way door.

4. Posted, immutable, and visible

How alerts fire

Alerts are evaluated the moment new data lands and whenever you open a dashboard. A fully autonomous background sweep — surfacing an alert with no one logged in — is on the near-term roadmap.

5. Out to your accountant and to QuickBooks