AI Agents for Business
Central AR
Redoing the 85 percent that already matched adds nothing. The 15 percent that did not is the whole opportunity.
- Business Central ships Copilot assisted bank reconciliation and it works. The residue it leaves is small in percentage terms and most of the working time.
- In a worked month, 62 unmatched lines worth 284,000 came down to five causes, and 44 were mechanical once the right references were read.
- Twelve lines a month came from one configuration gap where the customer's own reference never reached the invoice.
- Suppress anything not actually owed before producing a reminder list. On a mid market ledger the person chasing often knows the customer personally.
- Escalate on a change in payment behaviour rather than a level. A customer who always paid at day 45 is not news.
AI agents for Business Central accounts receivable work the residue that native bank reconciliation leaves behind, grouped by cause rather than listed, and build collections from observed payment behaviour rather than stated terms. ChatFin proposes applications and contacts nobody.
The starting point is that Copilot assisted reconciliation already does the easy part well, and this page does not argue with that. What it leaves is a payer name that matches nothing in the customer list, one receipt settling several invoices, a customer quoting their own reference, and short payments that turn out to be settlement discounts. That residue is where the afternoon goes.
Native bank reconciliation clears the easy lines. The residue is the job
Business Central ships Copilot assisted bank reconciliation and it works. What it leaves is the lines it could not resolve, and ChatFin finds that residue is small in percentage terms and most of the working time.
| Residue cause | Why native matching cannot resolve it |
|---|---|
| Payer name differs from the customer | The name on the statement is a parent or a bank |
| One receipt against many invoices | No single invoice matches the amount |
| Customer used their own reference | It is on your invoice only if billing captured it |
| Short by a settlement discount | The amount does not match and the difference is allowable |
| A genuine deduction | Needs a decision, not a match |
Redoing the 85 percent that already matched adds nothing. Working the 15 percent by cause is the whole opportunity, and ChatFin is built to start there rather than to replace what already works.
ChatFin works the residue and leaves native matching alone. Where Copilot already cleared a line there is no reason to touch it.
A worked residue, 62 unmatched lines
A month of statements leaves 62 lines unmatched, 284,000.
| Cause | Lines | Value | Resolution |
|---|---|---|---|
| Payer name differs from the customer | 19 | 96,000 | Resolvable from payment history |
| One receipt, many invoices | 14 | 118,000 | Combination match against the total |
| Customer reference not on the invoice | 12 | 41,000 | Billing setup. Recurs monthly |
| Settlement discount taken in terms | 11 | 19,000 | Allowable, apply on the rule |
| Genuine deduction | 6 | 10,000 | Commercial, not finance |
Forty four of the 62 are mechanical once the right references are read. Twelve recur every month until billing captures the customer's own reference, which is a configuration change rather than a monthly task.
Collections on a mid market ledger
Fewer customers, larger relative balances, and the person chasing often knows the customer personally. That changes the tone and the sequencing.
ChatFin rebuilds the open position immediately before any reminder list is produced and returns the suppression list with a reason for each. It contacts nobody.
Where this underperforms
Three.
Deeper in Business Central receivables
Three use cases sit under this hub, in build and shown rather than linked.
Worked on this page
AR inbox automation · Open AR analysis · Bank payment to open invoice matching
Covered as sections here
Bank reconciliation residue · Cash application · Collections · Credit notes · Deductions
Questions Business Central teams ask first
Business Central already reconciles the bank. What is left?
The lines it could not resolve, which is typically 15 percent of the statement and most of the working time. In a worked month, 62 unmatched lines came down to a payer name differing from the customer, one receipt against many invoices, a customer reference never captured on the invoice, settlement discounts and genuine deductions. ChatFin works that residue and leaves native matching alone.
What recurs every month and can be fixed once?
The customer reference that never reaches the invoice. In the worked example it produced 12 unmatched lines a month, and it is a billing configuration change rather than a monthly task. ChatFin reports the recurring causes separately from the one off ones, because only one of the two categories can be removed permanently.
How should collections work on a mid market ledger?
Suppress first, then rank. Anything covered by a credit, already paid or in dispute should not be on the list at all, because on a smaller ledger the person chasing often knows the customer personally and a wrong reminder costs more than the invoice. Then rank on observed payment behaviour, and escalate on a change rather than a level.
Work the 15 percent, not the 85 that already matched.
ChatFin works the bank reconciliation residue by cause, matching on the references customers actually used, and separates recurring causes from one off ones.
It rebuilds the open position before any reminder list and returns the suppression list with a reason for each entry.
Bring one month of unmatched bank lines. The comparison takes an hour.
Book a demoAll images are served from the Pexels content network. No third party product screenshots, vendor logos or licensed media are used on this page.
Bring one month of unmatched lines.
We will group them by cause through ChatFin and show you which ones will come back next month unless a setting changes.