The Collections Stack You Can Build Inside NetSuite

THE PROJECT

This summer another NetSuite firm brought me in to work on their own accounts receivable. The brief was rough guidelines around the functionality they wanted to see; the design was mine. What we ended up with is a pattern I think most mid-market NetSuite customers should see before they sign up for another subscription: a complete collections stack, built natively.

Four pieces. Here is what each one does and why it exists.

PIECE 1: DUNNING THAT RUNS ITSELF

The first two collection touches should never depend on someone remembering to send them. NetSuite's native dunning handles the reminder cadence: which customers get letters, at what aging thresholds, with what escalating tone, and when a human takes over from the automation.

The design decisions matter more than the feature. Which customers are excluded, because not every relationship should get an automated nudge. What the escalation path is after the letters stop working. Where automation ends and a phone call begins. Get those right and the automation buys your collectors hours every week without costing you a single relationship.

PIECE 2: THE WATCHLIST

An aging report tells you who is past due. It does not tell you who is drifting. The watchlist is a saved-search-driven view that surfaces the accounts actually trending toward trouble: balances growing quarter over quarter, payment timing stretching, credit exposure creeping toward the limit.

The difference in practice: instead of a collector scanning three hundred rows of aging every Monday, the watchlist hands them the short list that deserves attention this week. It turns collections from a report you read into a queue you work.

PIECE 3: DSO YOU CAN ACT ON

Everyone reports DSO. Very few teams report a DSO anyone can act on. A single blended number across the whole company hides more than it shows.

Broken out properly, by segment, by customer class, trended over time, DSO stops being a scorecard and starts being a diagnostic. A rising number in one segment while the blended figure holds steady is exactly the kind of signal a blended metric was hiding, and exactly what leadership actually wants to know.

PIECE 4: THE WORKBENCH

The person working the cash should not need seven tabs to work one account. The collections workbench is a single working view: the queue, the account history, the notes, the promises, the next action.

This is the piece that makes the other three real. Dunning, the watchlist, and DSO reporting all produce signals. The workbench is where a human turns signals into decisions, and it is the difference between tooling people admire and tooling people use.

HOW THE DESIGN WORK GOT DONE

One more thing worth saying plainly: I used AI as a design partner on this build. Working in Claude Cowork, I pressure-tested the dunning cadence and exclusion logic, argued through what earns a spot on the watchlist, and laid out the workbench flow before configuring anything. The decisions stayed mine. What the AI changed was the speed of getting to a defensible design, and how early the gaps surfaced on the whiteboard, where they are cheap, instead of in the build, where they are not.

THE AI LAYER

The reason this project happened now is that the AI in NetSuite has gotten good enough to be worth designing around. Payment predictions that read actual customer behavior can feed a watchlist directly: an account whose predicted pay dates are stretching belongs on the list before it ever hits an aging bucket. And the 2026.2 release is absorbing more of the mechanical work, cash application suggestions and reconciliation matching, which frees the humans for the judgment calls the workbench exists to support.

AI did not replace any part of this stack. It made each piece smarter.

WHEN AN ADD-ON ACTUALLY EARNS ITS PLACE

None of this means collections add-ons are a scam. High transaction volume, heavy deduction and chargeback management, specialized industry workflows: real use cases, and a good add-on pays for itself there.

But "we never configured what we already own" is not a use case. Before you evaluate a third-party collections tool, inventory what your NetSuite instance can already do with dunning, saved searches, and a thoughtfully built workbench. Price the configuration work honestly and compare it to the subscription, not just for year one but for year five.

The most expensive software in your stack is the tool you bought to avoid understanding the one you already had.