Use case
Reconciliation with Lucius
Definition
Reconciliation is the workflow that confirms every transaction is recorded, matched to the right source, and posted to the correct account, across bank activity, payment processors, invoices, payroll, and internal transfers. Traditional accounting reconstructs this at month-end; Lucius reconciles continuously on a stateful ledger. AI proposes matches with confidence scores; non-revenue transactions at 95% confidence or higher auto-approve, with human review for exceptions and all revenue matches.
The problem
- Bank, processor, and accounting systems each tell a different version of the truth, and finance teams spend the close reconciling differences.
- Manual matching is slow, error-prone, and breaks down at scale.
- Without an explainable trail, reconciliations are hard to audit and easy to lose context on.
How Lucius solves it
- Lucius reconciles bank activity, processor payouts, invoices, and internal records against the same maintained financial state.
- AI-assisted matching proposes high-confidence reconciliations for vendor, COA, invoice, and receivable. Humans approve, override, or split — the system learns from every decision.
- Every posting is event-linked and append-only, so the reconciliation trail is permanently explainable.
Workflow
- 1
Ingest activity
Bank, card, and processor activity sync into Lucius automatically.
- 2
Match against state
Lucius matches transactions against open invoices, vendor bills, payroll, and known patterns.
- 3
AI suggestion
High-confidence reconciliations are suggested with vendor, COA, invoice match, and VAT.
- 4
Approve, override, or split
Humans approve in bulk, override exceptions, or split transactions across accounts.
- 5
Post to stateful ledger
Approved reconciliations post journal entries that link back to the originating event.
Example data flow
| From | To | What happens |
|---|---|---|
| Bank / processor | Transaction | Raw transaction lands in Lucius with metadata and counterparties. |
| Transaction | Match candidates | Lucius scans open invoices, bills, payroll, and patterns for likely matches. |
| AI suggestion | Approved posting | Reviewer approves; Lucius posts double-entry journal entries linked to the source event. |
Frequently asked questions
What does reconciliation mean in Lucius?
Reconciliation in Lucius means matching every bank, card, and processor transaction to the right source event, an invoice, vendor bill, payroll line, or internal transfer, and posting to the stateful ledger with a complete audit trail. Lucius proposes matches with confidence scores; non-revenue transactions at 95% confidence or higher auto-approve, reserving human review for exceptions. The result is continuously reconciled books instead of a month-end cleanup cycle.
Does Lucius use AI for reconciliation?
Yes. Lucius uses AI to propose vendor, category, and invoice matches from transaction descriptions, amounts, and historical patterns. High-confidence suggestions auto-approve; revenue and edge cases route to experienced reviewers. Decisions become patterns the system learns, reducing manual work as volume scales without sacrificing control.
Can the ledger be edited after a reconciliation?
Corrections and reclassifications are append-only. Lucius posts new journal entries that reference the original posting rather than mutating history. That preserves a permanent audit trail required for venture-backed companies, tax filing, and board reporting, every change is explainable and traceable to a human decision.
Does Lucius reconcile Stripe payouts?
Yes. Stripe payouts, fees, refunds, and chargebacks are matched against the originating invoices and posted.