Accounts receivable is the last function in finance still running on instinct

CFO.com said it plainly this year: accounts receivable is the last function in finance still running on instinct. FP&A has models. Procurement has software. Payroll is automated. Treasury uses real-time dashboards. And AR? AR still largely runs on an analyst's judgment about which customers to call, which invoices to chase first, and whether that $40K overdue account from a customer the sales team "really cares about" is worth escalating today.

This isn't a criticism of AR teams—it's a structural reality. AR was never given the same tooling investment as other finance functions, and as a result, it operates more on relationship management and intuition than on data. For software companies, where the receivables book is often large, fast-growing, and full of complex billing arrangements, that gap is expensive.

What "running on instinct" actually costs

The cost shows up in a few places:

• Inconsistent follow-up: Without a defined escalation protocol, different AR analysts manage aging accounts differently. One sends reminders at day 15, day 30, and day 45. Another waits for someone to ask. The outcome is unpredictable, and the variance in your DSO reflects it.
• Relationship-sensitive accounts get undertreated: "Strategic" customers receive fewer escalations because someone on the AR team is worried about the sales relationship—even when those accounts are the largest and most delinquent. Instinct says protect the relationship; data would say this customer has been slow 4 of the last 6 months.
• Escalation timing is reactive, not triggered: Accounts get placed with external agencies after someone manually reviews the aged report and notices a problem, not because a system flagged an account at 90 days. The difference between proactive placement at 90 days and reactive placement at 180 days is 20–40 cents on the dollar in recovery odds.
• Dispute resolution lacks a documented process: When a customer disputes an invoice, the handling often depends on who picks it up. Documentation is scattered across inboxes. Outcomes aren't tracked in a way that builds institutional knowledge about which dispute responses work.

What changed in other finance functions—and why it hasn't happened in AR

FP&A went through its modernization cycle with tools like Adaptive Insights, Anaplan, and Pigment. Procurement got Coupa. Payroll got Workday. Even cash flow forecasting—historically a manual process—is now a software product category with multiple vendors offering real-time direct-method forecasting off live bank feeds.

AR modernization lagged for a few reasons. First, the output of AR work is "we got paid," which looks like a success even when the timing was bad and the process was inefficient. Second, AR sits at the intersection of finance, sales, and customer success—which makes ownership ambiguous and tooling decisions complicated. Third, the traditional AR software (billing, invoicing) was treated as a solved problem, so the automation layer—dunning intelligence, escalation workflows, payment prediction—never got prioritized.

That's changed. Platforms focused specifically on the AR automation problem (not billing, not invoicing—the follow-up and escalation workflow) have matured significantly in the last two years. And CFOs are paying attention because the ROI is direct and measurable.

What modern AR discipline actually looks like

The difference between an instinct-driven AR process and a disciplined one is visible in three places:

1. Defined escalation triggers, not analyst judgment. Every account has a sequence: day 15 reminder, day 30 formal notice, day 45 senior escalation, day 60 external referral inquiry, day 90 placement. The triggers are automatic. An analyst's judgment about whether to wait another week doesn't change them.

2. Payment behavior data, not payment history. Modern AR tools don't just track whether a customer paid—they track how they paid. Did they pay within 5 days of the reminder? Did they request an extension every cycle? Did their average days to pay increase over the last two quarters? Those patterns predict future behavior more accurately than a one-time late payment or a longstanding "good customer" designation.

3. External escalation as a defined process step, not a last resort. Companies with low bad debt rates treat external collection as a normal step in the AR workflow—something that happens at a specific trigger point, not something that gets considered only after months of internal attempts. The stigma around "sending to collections" is a cultural artifact, not a financial strategy. Placement at 90 days is a process discipline. Placement at 14 months is a scramble.

Where to start

If your AR process is still largely relationship-driven and manually managed, the best diagnostic question is this: pull your last 12 months of bad debt write-offs and calculate the average age of those accounts at the time they were written off. In most software companies, that number is 14–18 months. Ask yourself how many of those accounts were first escalated to an external agency, and at what age.

Most CFOs who do this exercise find that write-offs weren't inevitable—they were slow. The accounts were recoverable at month 6 or 8, but no one made a decision to escalate until month 14 or 15, by which point the debtor had restructured, changed personnel, or simply disappeared.

AR modernization doesn't require a six-month technology project. It starts with a written escalation policy, a defined external placement trigger, and a commitment to treating AR like the other finance functions that already run on data.