Building Stable Finance Processes in E-Commerce Through Automation: Here’s How

Invoicing, dunning, payment reconciliation: 3 steps in one process, not 3 separate tasks 

Something I’ve noticed at a lot of companies: they treat invoicing, dunning, and payment reconciliation as three separate jobs. That’s the first problem. In reality, the three steps are directly linked.

Whatever goes wrong in step 1 compounds in steps 2 and 3:

• Wrong invoice data (a wrong amount, a missing required field, an incorrect payment term) means the dunning process either doesn’t start at all or dun the wrong amount. The customer pushes back. The process stalls.
• A patchy dunning process (because someone has to decide manually who gets dunned) means some customers are never dunned, others too late. The resulting cash flow gaps don’t show up explicitly in any report. But they cost money every month.
• Manual payment reconciliation means the company doesn’t know day-by-day which invoices are paid. Reminders go to customers who paid days ago. Customer relationships suffer while the finance team loses hours on reconciliation the system could handle.

The goal is to set up each step so it automatically informs the next. That’s the core of a stable, automated finance process for growing companies.

As a benchmark for the finance processes of scaling midmarket businesses, I look at three properties the process needs to have:

1. Person-independent: The process runs even when the person who built it isn’t there.
2. Volume-independent: The process scales without 500 invoices requiring five times the work of 100.
3. Data-driven: Every step generates readable data (open items, payment status, dunning level) that can be used for strategic and operational decisions, not just bookkeeping.

Step 1: Invoicing as the base that carries everything else 

Automated invoicing with the ERP workflow "Invoice generation from order list" from the Xentral Workflow library works like this:

As soon as an order is completed and paid, the invoice is generated and sent automatically. No manual step needed, no waiting for someone to work through the list.

On top of that comes the "Set payment term on new orders" workflow. This means every invoice goes out with the customer-specific payment term rather than the default. Sounds like a detail, but it’s the difference between a dunning process running on correct data and one running on false premises from day one.

This should be obvious, but I’ll say it anyway: automated invoicing isn’t a comfort feature in e-commerce. It’s the prerequisite that lets steps 2 and 3 function at all.

Step 2: Automate dunning in the ERP - rule-based, not reactive 

For me, this is one of the biggest misconceptions about dunning: a lot of companies think they have a working dunning process because they send reminders when someone doesn’t pay. That’s only reactive dunning.

Automated dunning means: the dunning level triggers on rules when the payment term is exceeded. It should happen regardless of whether anyone is thinking about it.

A dunning process that needs a person to remember it isn’t a dunning process. It’s a plan that often doesn’t get executed.

What I typically see at companies: dunning levels exist on paper, but the exceptions pile up.

• “The customer called.” The reminder gets stopped manually, no follow-up.
• “The invoice is disputed.” The entire dunning process gets blocked for that customer, including other invoices.

Six months later, there are more exceptions than rules. The problem is that the exceptions aren’t cleanly documented and tracked.

What does this look like in Xentral, concretely? 

Xentral offers a fully automated dunning process with configurable dunning levels: send triggers, templates per level, custom deadlines. When the payment term is exceeded, the process runs without anyone intervening.

Exceptions can be set precisely. Individual invoices can be manually blocked from further dunning and land in the "Blocked" tab. That keeps oversight in place even when specific cases deviate.

The result: Automating dunning in the ERP doesn’t mean human decisions disappear. It means human decisions only happen when they’re needed. They’re not a prerequisite for the process to start.

Step 3: Payment reconciliation - the step that closes the loop 

Payment reconciliation is where the most manual hours vanish. At the same time, it’s the least prioritized step, because it usually “kind of” works.

Anyone reconciling weekly doesn’t know where they really stand. That’s not an exaggeration. If reconciliation isn’t current every day, the company doesn’t know which invoices are actually open.

Which means no reliable cash position. Decisions on purchasing, investment, or headcount are made on numbers that aren’t right.

Automating payment reconciliation in the ERP isn’t just about more certainty in accounting. Payment reconciliation is a liquidity decision.

What does this look like in Xentral, concretely? 

Automatic payment reconciliation runs directly in the payment intake module. The "Automatic reconciliation" button matches incoming payments against open items.

Incoming payments get reconciled against open items automatically in the ERP. Payment status updates in real time, with no manual searching through bank statements.

On top of this, we introduced a payment reconciliation agent, currently in beta. It handles the matching automatically and proposes the corresponding bookings. The agent prepares, I just decide. It recognizes partial payments and overpayments, proposes the fitting booking, and the finance team approves.

What I like is that the agent takes the approval principle seriously: automation where it removes work and human control where it’s needed.

Worth underlining: if you have the three building blocks running natively in one ERP, you’ve automatically laid the data foundation that AI agents in the ERP need to work meaningfully.

An agent working on fragmented data from three different tools is unreliable. Only when a payment reconciliation AI agent runs on a clean, central ERP data foundation can it match independently.

That’s the difference between an agent you trust and one you have to constantly check.

Checklist: what separates a stable from a fragile finance process?  

A stable finance process in the ERP isn’t a question of company size. I see companies with 5 million in revenue whose finance processes run more smoothly than at 50-million businesses. That’s because someone spent an hour setting the process up correctly. That hour pays off every month.

Here are the nine questions I ask any company that wants to check whether its finance process is really stable:

Invoicing

1. Does your invoicing (whether from Shopify, your shop system, or straight from the ERP) run automatically even when the responsible team member is out sick?
2. Does every invoice go out with the customer-specific payment term (instead of the default)?
3. Does your ERP invoice format automatically meet e-invoice requirements (ZUGFeRD, XRechnung) without a manual step in between?

Dunning

1. Does your system know which customers should be dunned without anyone checking?
2. Are exceptions (paused reminders) documented and traceable, or do they disappear without a note?
3. Does your automated dunning in the ERP run as reliably at 500 open items as at 50?

Payment reconciliation

1. Do you know today’s open-item balance in the ERP because automatic payment reconciliation runs hourly?
2. Are incoming payments automatically matched to open invoices, or is someone doing it by hand?
3. Does your dunning process automatically get the update when an invoice is paid, so no reminder goes out on a balance that’s already settled?

If you can answer yes to all nine, you have a stable finance process. Three or more nos means you have process gaps, not tool gaps.

Bottom line: setting up a stable finance process doesn’t take weeks 

It takes a few hours if the system cooperates. In Xentral ERP, everything runs through one system and automated: invoicing, dunning, payment reconciliation. The three steps mesh without a manual handoff. That doesn’t just make your finance team’s day easier. Finance automation also improves your company’s liquidity.

If you want to know what that looks like in your specific setup, we’re happy to show you.