# Automated Cash Application

Ledge's cash application agents match payments from banks, payment processors, and remittance files to invoices in your ERP in real time — so you always know what's been paid, what hasn't, and that everything has been applied correctly.

#### How it works

Cash application is a structured, recurring task that the agent can execute automatically based solely on the task details, without any user input.

From the checklist, open a cash application task (e.g., *Cash Reconciliation for Account 8475*). Click into the **Agent Studio** to see the agent's working paper.

#### Working paper structure

* **Bank Statement tab:** Raw data from the bank.
* **Cash Application tab:** The full detail of the agent's matching work, including:

| Column                 | What it shows                                            |
| ---------------------- | -------------------------------------------------------- |
| Payment info           | Bank reference, date, amount                             |
| Payer name (extracted) | Name extracted by the agent from the bank narrative      |
| Resolved entity name   | How the payer was matched to an entity in your ERP       |
| Match logic            | The reasoning the agent used for each match              |
| Match level            | The complexity tier of the match (see below)             |
| Invoice details        | Invoice reference, amount applied, any amount difference |
| Subsidiary             | The corresponding NetSuite subsidiary                    |
| Notes                  | Any additional context (e.g., "oldest invoice selected") |

#### Match levels

The agent matches payments starting from the most straightforward and working toward more complex cases:

* **Level 1 — Exact match:** Invoice reference and amount match exactly
* **Level 2 — Fuzzy name match:** Payer name and invoice entity are similar
* **Level 3 — Invoice reference match:** Names differ but invoice numbers align
* **Level 4 — Cross-entity match:** Payment from one entity matched to invoice from a related entity (flagged as cross-entity)
* The agent also handles: typos in your ERP, company rebranding/name changes, bank fees, and short payments or underpayments

#### Summary and key findings

The working paper includes summary statistics (match count, amount, percentage by level) and a text summary of key findings on the left-hand side. The agent flags any unmatched items and recommends next steps, such as:

* **Unmatched payment processor transactions** (e.g., ADN payouts) that need journal entries rather than invoice application
* **Cross-subsidiary payments** where intercompany routing may be needed
* **Bank fees** and **short payments**

#### Generating journal entries and posting

For unmatched transactions that require journal entries (e.g., payment processor payouts), you can instruct the agent directly in the chat interface to generate them. It will create all relevant journal entries — including extracting payout references from the processor — which you can then review and post directly to NetSuite.

For items the agent hasn't seen before, it can suggest how to handle them based on how similar transactions were treated in historical records.