Accounts Receivable Chaser with AI

# Accounts Receivable Chaser — SoloStack Skill File
> Drop this file into your SoloStack repo and Claude Code will use it as
> context to run this workflow automatically.
---
## Overview
Drafts follow-up emails for unpaid invoices with stage-aware tone — gentle nudge at day 3, firmer at day 14, escalation script at day 30. Lands in your review queue. Never auto-sends. The cheapest revenue you will ever recover.
## Capabilities
1. **Reads AR from the ledger** — Pulls every invoice with `status: pending` from the transaction ledger. Computes days since send, days past due, days since last contact.
2. **Stage-aware tone calibration** — Day 3: friendly nudge ('just making sure you got this'). Day 14: firmer ('payment was due — anything blocking?'). Day 30+: escalation script with options (payment plan, escalate to legal, write-off).
3. **Drafts in YOUR voice** — Reads past emails to the customer (if available) to match tone. Won't sound like a stock template the customer's inbox auto-filters.
4. **Review queue, never auto-send** — All drafts land in `departments/finance/ar-queue/`. You review, edit if needed, approve. Then sends via Resend. Never autonomous — too high-stakes.
5. **Logs every touch back to CRM** — Each chase email writes a CRM event (`category: 'ar-chase'`, `channel: 'email'`) so the customer's timeline is complete.
## Pipeline
## Use Cases
- **Weekly AR review:** Friday afternoon: run skill, get drafts for everything overdue, approve 5 in 10 minutes. Typical first-month recovery: $1k–10k of invoices that would have sat there indefinitely.
- **Stale invoice cleanup:** Quarterly: run with `--include-stale` to find invoices >60 days that need an escalation conversation, not just a nudge. Get the conversation-starter draft.
- **Cohort billing chase:** If you bill a cohort upfront and 3 out of 10 are unpaid by start date, run with `--cohort` to get parallel drafts in one batch.
## Warnings
## Sample Output
```json
{}
```
Download the complete skill file with full implementation details,
agent prompts, and ready-to-run scripts.