The 86/14 Rule: Why Your A/R Team Should Only Touch 14% of Collections

Most accounts receivable teams are drowning in work that does not need a human at all. In the US, 44% of B2B invoices are now overdue, and the manual friction of chasing them costs mid-market companies an estimated 2 to 5% of revenue every year (Atradius). Yet the bulk of that chasing is not skilled work. It is the same reminder, sent to the same kind of account, on the same schedule, over and over. The problem is not that your team is too small. It is that your team is spending its judgment on tasks that never required any. The fix is a simple split we call the 86/14 rule.

The 86/14 rule states that roughly 86% of collections work is repetitive, schedule-driven follow-up that an AI agent can run end to end, while about 14% is judgment work (real disputes, negotiations, and relationship calls) that should always stay with a person. In practice, automating the 86% and routing the 14% to humans with full context lets a team collect far more without adding headcount. One trades business working this way recovered $842K in a quarter and cut DSO 61%.

What Is the 86/14 Rule?

The 86/14 rule is a way to divide collections work by what actually needs human judgment. It rejects both extremes that dominate the conversation. It is not fire your A/R team and let a bot do everything, and it is not collections is too relationship-driven to automate. Both are wrong. The truth is that collections is two very different jobs wearing one title.

The first job is coverage: making sure every open invoice gets a timely, polite, multi-channel nudge on a fixed schedule. The second job is resolution: handling the accounts where something real is happening, a dispute, a hardship, a negotiation, a relationship worth protecting. Coverage is about 86% of the volume and needs consistency, not judgment. Resolution is about 14% and needs a person. The 86/14 rule says: automate the coverage, defend the resolution. Most teams fail because they make their best people do coverage, which means coverage slips and resolution gets rushed.

Why 86% of Collections Is Repetitive Work a Machine Should Do

The largest share of collections is not negotiation. It is logistics. And logistics is exactly what software does better than people.

The coverage bucket

This is the work that fills an A/R team's day and burns its time: sending the pre-due reminder, the on-due notice, and the structured past-due nudges; confirming receipt; resending the invoice the customer never got; chasing remittance details; and doing it across email, text, and phone for hundreds or thousands of accounts. None of it requires reading a situation. It requires showing up on time, every time, on every account. Humans are bad at that at scale, not because they lack skill but because attention is finite and the squeaky accounts always win. Machines do not get tired, distracted, or pulled into a fire drill, which is why automation lets the same team manage a much larger portfolio without new headcount (Gartner). The accounts that age are almost never the ones that refused to pay. They are the ones nobody got to, which is the same reason most A/R trouble is really a follow-up problem.

Why the 14% Is Exactly Where Your People Belong

The remaining sliver is where collections actually earns its reputation as relationship work, and where a human is non-negotiable.

The resolution bucket

This is the genuine dispute over scope or quality, the customer going through real hardship who needs a payment plan, the strategic account where the relationship is worth more than the invoice, and the negotiation that needs authority and empathy. These calls require judgment, context, and accountability. They are also where the money and the relationships are won or lost. The problem in a manual shop is that this 14% competes for time with the 86%, so your most capable collector spends the morning sending routine reminders and arrives at the hard conversation drained and behind. Pull the 86% off their plate and the 14% gets the attention it deserves. This is also why the human-in-the-loop design matters: an AI agent should escalate the moment an account stops being routine, handing a person the full history rather than a cold start. We go deeper on that handoff in can AI actually call your customers to collect.

How Do You Tell the 86 from the 14?

The split is not a guess. You can sort almost any account with one question: does this need a decision, or does it just need a touch?

• If the account is simply unpaid and not yet contacted enough, it is coverage. Automate it.

• If the customer has raised a dispute, asked for terms, or signaled hardship, it is resolution. Route it to a person.

• If an account has had consistent follow-up and still will not engage, it is a judgment call about escalation, which is resolution.

• If you are sending the same message you have sent a hundred times, it is coverage, no matter how important the customer.

The mistake teams make is treating account size as the sorting line, putting every big customer in the human pile. Size is not the test. A large account that just needs a reminder is still coverage. A small account in genuine dispute is resolution. Sort by judgment required, not dollars at stake.

What the 86/14 Split Does to Your Numbers

Done right, the model moves three numbers at once.

DSO drops, because coverage stops slipping

Most of the gap between your DSO and your industry benchmark is uncontacted or under-contacted accounts aging quietly. When coverage runs automatically on every invoice, that drift disappears. The trades client above cut DSO 61% this way.

Capacity multiplies, without hiring

When the 86% runs itself, the same headcount can cover far more receivables. You are not replacing the team. You are removing the work that was capping how much they could handle, so a small A/R function can run a book that used to need three more hires.

Cost per dollar collected falls

Manual follow-up is the expensive part of A/R, the 2 to 5% of revenue lost to friction. Automating coverage takes the largest cost driver and makes it close to fixed, while your people spend their (expensive) time only on the 14% where judgment pays for itself.

How to Implement the 86/14 Model

You do not need to rebuild your stack to start. The model is mostly about reassigning work.

First, audit a month of your collections activity and tag each action as coverage or resolution. Most teams are shocked to find 80%+ of their effort is coverage. Second, define your escalation triggers in writing: the exact signals (dispute raised, plan requested, no engagement after a full cadence) that move an account from the machine to a person. Third, automate the coverage on a fixed, multi-channel cadence so no invoice ages for lack of a touch. Fourth, route the 14% to your team with full context attached, never a cold handoff.

This is exactly how Abivo is built. An AI agent runs the 86% across call, text, and email on your existing systems, and escalates the 14% to your people with the full account history. Most teams are live in under a week, and the platform is SOC 2, ISO 27001, PIPEDA, and HIPAA compliant. You can see how this plays out in our case studies. The framing matters more than the vendor though: even if you never automate a thing, sorting your collections into coverage and resolution will tell you where your team is wasting its judgment.

This is also the direction the whole function is heading. Gartner expects that by 2029, one-third of finance staff will hold shared jobs where AI and a person jointly own a single role. The 86/14 rule is just that future, written down early for collections.

Practical Takeaways

• Collections is two jobs in one title: coverage (about 86%, repetitive follow-up) and resolution (about 14%, judgment work). Automate the first, defend the second.

• Sort accounts by judgment required, not dollar size. A big account that needs a reminder is still coverage.

• The accounts that age are usually the ones nobody got to, not the ones that refused. Consistent coverage fixes that.

• Write your escalation triggers down so the handoff from machine to human is clean and never a cold start.

• Done right, the split lowers DSO, multiplies capacity without hiring, and cuts cost per dollar collected.

Frequently Asked Questions

What is the 86/14 rule in collections?

It is a framework that splits collections by judgment required: roughly 86% is repetitive, schedule-driven follow-up an AI agent can run end to end, and about 14% is judgment work (disputes, negotiations, hardship, key relationships) that should stay with a person. You automate the 86% and route the 14% to humans with full context.

Does the 86/14 rule mean replacing my A/R team?

No. It means removing the repetitive coverage work that caps how much your team can handle, so the same people manage a larger book and spend their time only on accounts that need judgment. It is human-in-the-loop by design, not human-out.

How do I know which accounts to automate?

Ask whether the account needs a decision or just a touch. Unpaid-but-not-yet-worked accounts are coverage and should be automated. Anything with a dispute, a request for terms, hardship, or repeated non-engagement is resolution and should go to a person.

Where does the 86% number come from?

It reflects what Abivo sees across live collections: about 86% of A/R outreach is routine follow-up handled autonomously, and roughly 14% needs a human. Your exact split will vary, but most teams find the overwhelming majority of their collections effort is repetitive coverage.

What results does the 86/14 model produce?

By making coverage consistent instead of slipping, it lowers DSO, lets a team manage far more receivables without new hires, and cuts the cost of collections. One trades client recovered $842K in a quarter and cut DSO 61% running this way.

Tag one month of your collections activity as coverage or resolution. If the coverage pile is as big as it is for most teams, your people are spending judgment on work that never needed it. To see what automating the 86% does to your DSO and capacity, hit Get Started and we will show you.