AI Bookkeeping and Collections for Plumbing Businesses

Most plumbing shops close their books on the 15th of the following month. The good ones close on the 5th. The difference is not the bookkeeper's competence — it is whether the reconciliation between ServiceTitan (or Housecall Pro, or FieldEdge) and QuickBooks is automated or manual.

AI does not replace the bookkeeper. It replaces the rote reconciliation work, automates the dunning sequence on aged invoices, and surfaces the categorization exceptions the bookkeeper actually needs to look at. The result for an 8-truck plumbing shop is books closed five days faster, DSO compressed from 47 days to 28, and bad debt cut by 40–60%.

The four-part AI bookkeeping workflow

1. FSM-to-QuickBooks reconciliation

Every plumbing shop runs two systems of record. The FSM owns the job, the invoice, the payment record. QuickBooks owns the GL, the bank, the tax return. The gap between them is where errors hide.

AI reconciliation reads the daily FSM revenue, matches it line-by-line against QuickBooks deposits, flags variance over $50, and posts a clean journal entry. The bookkeeper reviews the exceptions instead of building the entry from scratch. Daily reconciliation that used to take 90 minutes runs in 15.

2. Automated dunning

The aged-receivables problem is not that customers do not pay. It is that nobody owns the follow-up past day 7. AI runs a 5-touch dunning cadence across SMS, email, and outbound voice:

• Day 7. Friendly SMS reminder with payment link.
• Day 14. Email with statement attached.
• Day 21. Outbound voice call with auto-drafted script.
• Day 30. Escalation SMS plus second voice call.
• Day 45. Final notice email; flag for collections decision.

Every interaction posts back to the FSM. The office manager only sees the cases that escalate. Most shops compress DSO from 47 days to 28 inside one quarter.

3. Expense and bill categorization

AI reads incoming bills (parts vendor invoices, fleet expenses, utility bills), matches against the chart of accounts, and posts a draft journal entry. The bookkeeper approves in batch. Errors that used to surface during month-end close get caught daily.

4. Cash-flow forecasting

AI builds a 13-week cash forecast from FSM AR, FSM AP, scheduled job pipeline, and historical seasonality. The owner sees the forecast on Monday morning. Most shops we work with use this to time payroll, sub-vendor payments, and capex against the actual cash position instead of a gut feel.

Vendor and integration

Three credible buyer paths:

• Native FSM accounting bridge. ServiceTitan's QuickBooks Online integration, Housecall Pro's QBO sync, FieldEdge's accounting bridge. Turn these on first. The native integrations cover 60–75% of the reconciliation work for free.
• Bookkeeping-specific AI. Bill.com with AP automation, Vic.ai for invoice processing, Truewind or Digits for AI bookkeeping overlays. $200–$800/month per location.
• Custom dunning automation. A Make.com or Zapier workflow that reads aged invoices from the FSM and runs SMS/email sequences. Build cost runs $2,000–$5,000 one-time.

For most 5–25 truck plumbing shops, the right stack is: FSM native sync + Bill.com for AP + a Numa or Podium dunning sequence for AR.

The 9-day rollout

1. Days 1–2. Audit. Pull last 3 months of GL entries, FSM revenue reports, and AR aging. Identify variance categories — are the daily entries off by deposit timing, refunds, partial payments, or coding errors?
2. Day 3. Configure the FSM-to-QBO sync. Map deposit accounts, refund accounts, tax-line categories. Run a one-day sync test.
3. Days 4–5. Build the dunning cadence. Write SMS, email, and voice scripts for each touch. Configure the FSM payment-link integration.
4. Day 6. Configure AP automation. Connect Bill.com or Vic.ai to the most active parts vendors. Train the category rules.
5. Day 7. Shadow week. Bookkeeper reviews every AI-generated entry. Catches edge cases.
6. Day 8. Live. AI runs the daily reconciliation, the dunning cadence, and the AP categorization.
7. Day 9. Measure. Time-to-close (target: 5 business days), DSO trajectory, dunning response rate.

What good looks like

• Time to close. Floor: 15 business days. Target: 5–7.
• DSO. Floor: 47 days. Target: 28–32.
• Bad-debt write-off. Floor: 4–6% of AR. Target: 1.5–2.5%.
• Bookkeeper hours per week. Floor: 18–24. Target: 9–12 (redirected to FP&A and exception handling).

Track in the monthly close packet alongside the job profitability by tech report.

Pitfalls

Letting AI auto-post without review. For the first 90 days, every AI-generated entry should be reviewed before posting. After that, low-risk categories (utility bills, recurring vendor invoices) can auto-post. High-risk categories (refunds, sales tax adjustments, owner draws) stay manual forever.

Aggressive dunning tone. AI-drafted dunning copy should escalate in firmness, not in hostility. Day 45 should be firm and procedural, not threatening. The customer who pays after day 30 is still a customer; preserve the relationship.

Skipping payment-link integration. Dunning without a one-click payment link converts at 12%. Dunning with a one-click payment link converts at 38%. Build the link.

Wrong chart of accounts. AI cannot rescue a chart of accounts that has 200 redundant categories. Clean it before turning on AP automation.

Ignoring sales tax. Most plumbing service work is taxable in most states; some material-only work is not. The AI must distinguish. Code the rule set or the books will be off at year-end.

How it fits the broader stack

AI bookkeeping is the cash-flow line in a plumbing AI rollout. The receptionist captures revenue, the dispatcher routes it, the estimating engine closes it, the dunning workflow collects it. Sequencing lives in the 2026 plumbing AI playbook.

FAQ

Q: Will my bookkeeper still have a job? A: Yes. The bookkeeper moves from data-entry to FP&A and exception handling. The ones we have worked with end up doing higher-value work and stay longer.

Q: What about my CPA? A: Your CPA closes the year, files the return, and advises on entity structure. AI bookkeeping makes their job easier and cheaper — most CPAs charge less for shops with clean books.

Q: Can the AI handle multi-entity shops? A: Yes, with configuration. If you have a service entity and a real-estate entity, map the chart of accounts carefully and configure inter-company rules.

Q: How does this handle credit-card surcharges? A: Configure the surcharge rule per state. Some states allow plumbing shops to pass card processing fees; some do not. The AI reads the rule and posts correctly.

Q: What about tech commissions and spiffs? A: AI handles standard commission calculations from FSM revenue data. Custom spiff programs (manufacturer rebates, install bonuses) usually need manual setup but auto-calculate after that.

Q: How does this fit with AI enablement generally? A: Bookkeeping AI is the back-office workflow that closes the loop between revenue and cash. It is the fifth of six standard workflows we sequence.

---

If you want a sized look at what AI bookkeeping would change in your close speed and DSO — your current cadence, your aging report, your monthly hours — reach out. Or see the full engagement on the AI for plumbers page.

SOURCES

Cited and consulted.

1. 01ServiceTitan Blog — Accounting and QuickBooks Integrationservicetitan.com · accessed May 8, 2026
2. 02Housecall Pro Blog — Bookkeeping for Home Serviceshousecallpro.com · accessed May 8, 2026
3. 03Plumbing & Mechanical Magazine — Financial Performancepmmag.com · accessed May 8, 2026
4. 04MSCA — Financial Benchmarking Researchmscanet.org · accessed May 8, 2026