Every AR manager I've talked to has a version of the same story: they called a good customer about an overdue invoice, and the customer was surprised — they thought it was paid. The relationship got awkward for a week. The payment came in two days later anyway. That call was a waste of everyone's time and a small withdrawal from the relationship bank.
The flip side: they spent three weeks sending emails to an account that had no intention of paying without a direct conversation. Each email got ignored. A 10-minute phone call on day 15 would have resolved it.
The escalation decision — when to move from email to phone, and which accounts to prioritize for human contact — is one of the highest-leverage judgment calls in B2B collections. Getting it wrong doesn't just cost you time. It costs you money in delayed payments, and occasionally it costs you relationships.
The Default Approach and Why It Fails
Most AR teams operate with a time-based escalation rule: send email reminders at net+5, net+15, net+30, then call at net+45. The logic seems reasonable — give the customer time to process, escalate if they don't respond. In practice, this fails on both ends of the spectrum.
It's too slow for genuinely at-risk accounts. A customer who goes completely silent after the first email, has a history of disputes, and has a $120K balance outstanding doesn't need to sit in the email queue until day 45. That account needed a phone call by day 20.
It's also too aggressive for administratively slow payers. Some customers consistently pay between net+35 and net+42. They've been doing it for three years. Getting a phone call on day 45 — which their AP team receives as a pressure call — doesn't accelerate payment. It just creates friction. You've interrupted someone's workday to collect a payment that was coming next week regardless.
The problem with uniform time-based escalation is that it treats all overdue accounts as the same kind of problem. They're not. Some are slow but reliable. Some are in internal dispute. Some have cash flow issues. And some are actively avoiding the invoice. The right escalation response is different for each.
The Signals That Actually Predict Escalation Need
After working through a lot of AR data, the escalation signals that actually matter sort into three categories: behavior change, relationship signals, and account characteristics.
Behavior Change
This is the most reliable escalation trigger. A customer who normally pays at net+28 but is now at net+40 with no contact is a very different signal than a customer whose historical average is net+38 and is currently at net+40. The first is a deviation from pattern. The second is normal variance.
Escalation-worthy behavior changes include: an account that has opened your emails but not responded (they've seen it, something is blocking), an account that previously always acknowledged receipt but has gone silent, or an account whose payment timing has lengthened by 15+ days compared to their 12-month average. These are signals of a change in the account's situation — not just administrative latency.
Relationship Signals
High-value, long-term customers warrant earlier escalation — not to pressure them, but because these are relationships where a direct check-in is appropriate. A customer who has done $800K with you in the last two years deserves a call from a human when something looks off. They don't deserve to receive three templated reminder emails before anyone picks up the phone.
Conversely, transactional customers with no significant history warrant less relationship-sensitive escalation. The calculus is different. You're not protecting a $2M pipeline; you're collecting a $4,500 invoice from an account that may not reorder.
Account Characteristics
Dollar amount matters, but not in the obvious way. It's not just about calling the big invoices — it's about calling the invoices where the ratio of relationship value to invoice amount justifies human contact. A $50K invoice from a $500K-per-year account has different escalation logic than a $50K invoice from an account you've invoiced once.
Dispute history also predicts escalation need. Accounts with prior short-pays or formal disputes are higher escalation candidates because their issues don't resolve through email. They need conversation to unblock.
A Practical Escalation Framework
We think about escalation across two dimensions: urgency and relationship sensitivity. These combine into four quadrants, each with a different contact strategy.
High urgency + high relationship value: Call early, prepare context before calling, frame as a check-in not a collection call. "We noticed invoice #4821 from July hasn't cleared yet — wanted to make sure everything was received correctly." This surfaces issues (disputes, routing problems, internal approval delays) that email alone won't reveal.
High urgency + lower relationship value: Firm email escalation sequence, then call at 21–30 days past due. No ambiguity about the purpose of contact. These accounts need clarity about your collections process, not relationship management.
Lower urgency + high relationship value: Standard email sequence but flag for personal review at each step. These accounts probably resolve themselves. But if they don't, a human call is the right recovery mechanism — not escalating to a collections-focused tone in email.
Lower urgency + lower relationship value: Full automated sequence, escalate to phone only if no response after 4–5 emails and amount exceeds your minimum collection threshold. Most of these resolve. The ones that don't are often uneconomical to pursue aggressively.
What Good Escalation Calls Actually Sound Like
The AR managers who get the best results from escalation calls share a few habits. They call with context. Before picking up the phone, they've reviewed the invoice number, the amount, the email history, any prior disputes, and the contact's name. The call takes 90 seconds of prep and the difference in outcome is significant.
They open with a question, not a statement. "I wanted to check in on invoice #4821 — has everything on your end been resolved?" leaves room for the account to surface a problem (wrong PO reference, internal routing issue, a dispute you didn't know about). Opening with "I'm calling about your overdue balance" closes that door.
And they document the outcome immediately. If the call produces a payment commitment ("we'll process it by Thursday"), that date goes in the system. The account gets a follow-up email confirming the commitment. If Thursday passes without payment, escalation resumes with documented context — "as discussed on September 22nd" carries weight that a generic reminder doesn't.
The Role of Automation in Escalation
We're not saying automation replaces the escalation call. It doesn't, and any tool that claims to fully automate collections is describing a different product category (debt collection software, which operates under different rules and serves a different use case).
What automation can do is dramatically improve the inputs to the escalation decision. If your system tracks email opens, payment history, behavioral deviations, dispute patterns, and account value — and surfaces a prioritized escalation queue with that context attached — your AR team makes better calls faster. They're not spending 25 calls making judgment calls that could have been automated away. They're making 8 calls on the accounts that genuinely need them, with full context, at the right time.
The human call is irreplaceable for certain accounts. The decision about which accounts those are — that's where good process and good tooling make the difference.