November 19, 2026

THE ACCOUNTING DEPARTMENT OF THE FUTURE SHOULD PAY FOR ITSELF

Topics:

Imagine opening Slack (or email) at 7:00 a.m. and seeing this message:

Books through yesterday are closed. Four exceptions need review. Cash is at X. Accounts receivable changed by Y. One unusual payment needs approval.

Nobody spent the morning rebuilding a spreadsheet. The system pulled information from the bank, cards, payroll, billing, and accounting file. Rules handled the routine work. AI helped compare the activity and explain what changed. A person still owns the judgment calls.

That is the accounting department of the future.

It does not wait until the 15th of the month to tell the owner what happened. It gives the business a clearer view while there is still time to make a different decision.

The goal is bigger than faster bookkeeping. The future accounting department should find money, protect margin, improve cash timing, and create capacity. It should pay for itself through the value it helps the business recover.

THE ACCOUNTING DEPARTMENT OF THE FUTURE

A business we worked with contract worth about $3 million had roughly $400,000 of work that had not been billed. The owner already suspected something was off. The gap became visible only after someone pulled the contract, job progress, costs, and billing together.

The business had done the work. The costs had started moving. The cash had not.

That is a simple test for the accounting department of the future. Can it connect the work the business completed to the money the business should collect? Can it surface the gap while someone can still send the invoice, resolve the issue, and bring the cash forward?

The answer requires more than an accounting file. You need the contract, delivery data, job costs, billing status, and customer terms all have to be available in the same operating picture. Today, for most businesses they’re not. But with AI, this becomes a reality.

The system can then compare what should have happened with what actually happened. AI can help explain the difference and route the exception and a person still decides what to do.

That is how accounting becomes part of the operating system of the business. It helps the owner see where money is stuck before the problem shows up as a cash shortage or a margin surprise.

ONE LOOP, THREE MOVEMENTS

The same idea applies to almost every important finance process. Think about a 13-week cash flow forecast. For one business, producing the forecast used to take a bookkeeper several hours each month. Once the banking and accounting data were connected with rules and scripts, much of the report became auto-generated.

The team still reviewed the forecast, but their time shifted from assembling the report to investigating the differences between what the business expected and what actually happened. The finance team gets more time for judgment because the system handles more of the movement between raw information and a decision-ready view.

Every process in the future accounting department has to make three movements.

First, the data has to move. Information from the bank, payroll, billing, cards, job system, and accounting file needs a dependable path into the finance operation. The data can live in different tools. The business needs a consistent way to bring the important pieces together and understand where each number came from.

Second, the work has to move. Someone needs to see the process from the first input to the final decision. Where does a bill enter? Who approves it? What happens when the amount does not match the purchase order? When does a completed job become an invoice? What happens when the customer disputes it?

Third, the judgment has to move to the right person. Routine items can follow rules. AI can compare, classify, summarize, draft, and flag. Exceptions that carry financial or operational risk need a human owner. The process should make that handoff visible instead of burying it in an inbox or in someone's memory.

Those three movements are the thread running through this series. Data access makes the information reachable. Process mapping makes the work visible. Sorting the work between rules, AI, and people makes the system safe enough to use every day.

WHY THIS IS HARDER THAN ADDING AI

Most accounting departments were built for periodic work by hand. Someone downloads a file, copies information into another system, checks a few totals, and rebuilds the report again next month. The process works until volume increases, a system changes, or the one person who knows the steps is unavailable.

Important data may already exist in a bank feed, PDF, inbox, spreadsheet, payroll system, billing tool, or job platform. The business owns the information, but nobody has built a dependable way to bring it together. A model cannot answer a question with data it cannot reach.

The process may also be unclear. One person knows which report to run, which exceptions matter, and who needs to approve the next step. That knowledge disappears from the business when the person is unavailable. Automation applied to an unclear process only makes the confusion travel faster.

This is why the future accounting department needs more than a new tool. It needs a reliable data path, an understandable process, and clear ownership for the decisions that carry risk.

THE VALUE YOU CAN FIND FIRST

You can test this without rebuilding the whole finance function. Pick one recurring loop: billing, collections, payables, payroll, month-end close, or cash forecasting.

Start with the first question the process is supposed to answer. For billing, that might be: what work is complete but not invoiced? For cash forecasting, it might be: what cash is expected to arrive and leave over the next 13 weeks? For payables, it might be: which payments are routine and which ones need an owner to review them?

Then trace the loop from input to decision. Write down every system involved, every handoff, every approval, and every place someone copies or searches for information. Mark the step where the work slows down or depends on one person's memory.

Finally, estimate the cost of the delay. A missed invoice has a dollar value. A late forecast has a decision cost. A manual report has a labor cost. A process that stops when someone is away has a capacity cost.

That gives you a starting point for the first improvement. Connect one missing data source. Document one handoff. Automate one repeatable step. Route one type of exception to the person who can act on it.

The first project does not need to be impressive. It needs to prove that the finance operation can move information from the source to a better decision with less friction.

WHAT WE WILL BUILD NEXT

This article is the introduction to a series about building that kind of accounting department. We will start with data access because every useful answer depends on the business being able to reach the right information. Then we will map the work that produces the numbers, including the handoffs, delays, approvals, and single-person dependencies that make finance fragile.

From there, we will sort the work. Rules should handle the predictable steps. AI should help with comparison, context, and recommendations. People should own exceptions, approvals, controls, and decisions that require judgment.

Over the next 6 to 8 weeks, each article will take one part of the future department and make it practical. You will see what the system is supposed to do, where teams usually get stuck, and how to choose a first process that can show value quickly.

In the next 24 hours, choose one finance loop and answer four questions: what starts it, what information does it need, where does it slow down, and who owns the next decision?

That is enough to begin. The accounting department of the future will be built one connected loop at a time.

I can’t wait to walk through this with you! We’ll take next week off (happy thanksgiving!) and get started December 3. See you then.

‍