| name | ar-collections |
| description | Use when reviewing accounts receivable aging, prioritizing overdue invoices, preparing collector worklists, identifying disputed balances, and recommending next collection actions for finance operators. |
AR Collections
Use this skill to turn an AR aging report and account context into a prioritized, source-backed collections plan.
Inputs
Ask for the missing items only when they are necessary:
- AR aging export with customer, invoice, due date, amount, currency, aging bucket, and owner.
- Customer notes, payment promises, dispute status, and prior outreach.
- Materiality threshold, collection policy, and escalation rules.
- Customer health context, contract status, open support tickets, or renewal timing.
Workflow
- Normalize the aging view by customer, invoice, due date, amount, aging bucket, and dispute status.
- Separate true collection work from administrative cleanup: unapplied cash, credits, duplicate invoices, invalid contacts, and invoices already under dispute.
- Rank accounts by action priority using amount, days overdue, relationship risk, dispute status, payment history, and upcoming renewal or churn risk.
- Assign a recommended next action for each account: reminder, statement resend, dispute resolution, sales/customer-success escalation, executive escalation, payment plan, write-off review, or no action.
- Flag missing evidence and assumptions. Do not invent payment commitments or customer history.
Output
Return:
- Executive summary: total past due, material exposure, urgent accounts, and main blockers.
- Priority table with customer, invoice count, past-due amount, oldest due date, status, owner, next action, and evidence.
- Dispute and cleanup queue.
- Outreach drafts only when requested.
- Approval list for sensitive actions such as executive escalation, service hold, payment plan, write-off, or legal referral.
Guardrails
- Do not threaten suspension, legal action, or credit reporting unless the user's policy explicitly allows it.
- Do not mark invoices as collectible if source data shows an active dispute, billing error, or unapplied payment.
- Keep customer-facing language factual and calm.
- Clearly distinguish recommendations from actions already taken.