AI controller vs fractional controller: which does your firm need?

The short answer: when you hire, hire for the work only a person can do, not for daily controller work software can carry. Whether you need a fractional controller depends on what’s actually missing in your firm.

• Your people have the judgement and the accounting is handled, but daily controller work eats their time. Start with an AI controller. Hiring another person to absorb that work is often the expensive way to solve it.
• Nobody in your firm owns ongoing financial review and recommendations for clients. Add a qualified person, a fractional controller or fractional CFO, and pair them with an AI controller.
• The expertise is there, but the human work still exceeds your people’s capacity after the daily work is automated. Another person may still be the right call.

The comparison sounds like software against a person. It’s really a question about which kind of work your firm is short on. This guide separates the kinds of controller work, shows which an AI controller performs, and helps you tell which situation you’re in.

What is a fractional controller?

A fractional controller is a controller hired part-time rather than full-time. A firm or a business brings one on for a set number of hours a week or a monthly retainer, instead of employing a salaried controller. The fractional controller might work independently or through a firm that supplies controller services, onshore or offshore.

Their scope varies, and it often goes well beyond review: closing the books, resolving complex accounting issues, preparing financial statements, and overseeing the accounting operation. You are buying controller-level work by the hour, so as the client book grows, the hours grow with it. Our guide on how many clients a fractional controller can carry looks at where that ceiling sits.

What is an AI controller?

An AI controller does a controller’s daily work across a firm’s client book. It is software that watches each client’s bank and ledger every day, keeps the forecast current, checks the plan against cash, and carries what was decided into the next call. The firm keeps the judgement. What is an AI controller? covers the definition in full.

The key difference isn’t human versus software. It’s that an AI controller does this work for every client, every day, and its capacity doesn’t shrink as the client book grows.

One premise before comparing

This comparison assumes the accounting itself is already covered: bookkeeping, reconciliation and closing the books each month are owned by someone, whether in your firm or the client’s. An AI controller works from those records. It doesn’t keep them.

If closing the books, resolving complex accounting issues or preparing financial statements is the work you’re short on, you need accounting capacity, and a fractional controller who owns that scope is a natural fit. The rest of this guide is about the controller work that sits on top of the books once they exist.

What controller work an AI controller performs

Controller work on top of the books falls into two kinds.

Daily controller work. This is the continuous, analytical part of the role:

• monitoring cash against what was expected, every day
• keeping the short-term forecast current as new activity lands, and grading it against what the bank later shows
• testing scenarios, like what a hire, a loan or a delayed payment would do to cash
• checking actuals against budgets and targets
• explaining what changed since the last conversation, and why
• tracking what was agreed on each call against what actually happened
• preparing the brief before each client call, and answering client questions in between

This is substantive work, and it’s the work an AI controller performs. It runs for every client, every day, without the hours climbing as the book grows.

Judgement and accounting resolution. This is the part that needs a person: deciding what the numbers mean for this business, recommending what to do, standing behind that recommendation with the client, and doing the accounting work to fix a problem once it’s found.

That last point is worth being precise about. An AI controller can detect a problem: a payment that didn’t arrive, a forecast that’s drifting, a record that doesn’t match the bank. Resolving it, whether that means correcting an entry, reworking an accrual or restating a month, is accounting work, and it belongs to a person.

Here’s what often happens without an AI controller. A firm hires a fractional controller for their judgement and accounting depth, and a large share of their hours goes to the daily controller work instead: rebuilding the forecast, checking the bank, assembling the picture before each conversation. That isn’t anyone’s fault. The work has to happen, and until recently a person was the only way to do it. But it means expensive expertise spends much of its week on work that can now be carried by software.

Where controller work sits in client accounting services

The profession already has a useful way to describe this. Client accounting services are commonly described as a ladder with three rungs:

1. Bookkeeping: recording transactions and producing the records.
2. Controllership: making sure those records can be relied on, and keeping the forward view current between closes.
3. CFO-level advisory: strategic judgement about where the business is going and what it should do.

The controller rung sits in the middle, connecting the records below to the judgement above. What a controller does that a bookkeeper does not explains the line between the first two rungs.

An AI controller performs the daily work of that middle rung. Whether you also need a person depends on what else your firm is short on.

If your firm has the judgement, but daily controller work eats its time

This is the shape of many CPA firms doing ongoing client advisory work, and firms with CFOs or fractional CFOs on staff. The accounting is handled. Someone already reads the numbers, forms a view and advises the client on a regular cadence.

What’s stretched is time. The partner or CFO spends hours before each client call rebuilding the picture, and across twenty or thirty clients there’s no room to stay on top of each one between calls.

In that situation, hiring a fractional controller to absorb the load is often the expensive answer to the wrong problem. The firm isn’t short on judgement or accounting skill. It’s short on the hours the daily controller work consumes. An AI controller performs that work across the whole client book, so the firm’s own people get their time back for judgement and client conversations.

AI controller for CPA firms covers how a small firm runs one in practice, from setup to what the firm still owns.

If nobody in your firm owns financial review and recommendations

This is the shape of a bookkeeping service, or a firm focused on tax and compliance where ongoing financial review and recommendations for clients aren’t part of the service today. The expertise may exist in the firm, but nobody is responsible for reading each client’s numbers between filings and advising on what to do next.

Here an AI controller isn’t the whole answer. It performs the daily controller work, but someone has to own the review and the recommendations above it, and software shouldn’t be the one to do that. So add a qualified person: a fractional controller or a fractional CFO.

This is where the two work best together. The person does what you hired them for: the judgement, the accounting depth, the client conversations. The AI controller performs the daily controller work, so their hours don’t disappear into rebuilding forecasts and assembling the picture before each call. You get more judgement from the same hours, and as the client book grows, the daily work doesn’t demand a matching rise in hours.

If the expertise is there, but the work still exceeds capacity

Sometimes a firm has the judgement and the accounting skill, automates the daily controller work, and still has more human work than people. More client conversations than anyone can hold. More accounting issues to resolve than the team can reach. More reviews than the partners have hours for.

That’s a genuine capacity gap, and another person may be the right answer. The point of an AI controller isn’t to argue you never need to hire. It’s to make sure that when you do hire, it’s for the work only a person can do, not for daily controller work software can carry.

How to tell which situation you’re in

Four questions usually settle it:

1. Is the accounting covered? If bookkeeping, reconciliation and closing the books are owned and current, the comparison applies. If not, that’s the gap to fill first.
2. Does someone in your firm already own financial review and recommendations for clients? Not just at tax time, but ongoing, forward-looking conversations about cash, plans and decisions. If not, you need a qualified person.
3. Where do your senior people’s hours go today? If much of their time goes to rebuilding forecasts, checking the bank and assembling the picture before each call, the gap is daily controller work, and an AI controller performs it.
4. Once that daily work is automated, is there still more human work than people? If there is, after the daily work is covered, that’s the case for another hire.

What neither one does alone

An AI controller doesn’t own the judgement, and it doesn’t do the accounting work to fix what it finds. It shows what changed, what was expected and whether it happened, models options with its assumptions visible, and flags problems early. A person reviews the recommendation, and the client approves the decision. What an AI CFO can’t sign off on explains why accountability can’t sit with software.

A person, on the other hand, can’t perform daily controller work for every client every day. Their hours are finite, and that work grows with every client added. That’s the structural limit an AI controller exists to lift.

The answer for most firms isn’t simply one or the other. It’s making sure the daily controller work runs on software that doesn’t run out of hours, and that the people you hire are hired for the work only people can do.

Frequently asked questions

What is the difference between an AI controller and a fractional controller?

A fractional controller is a person hired part-time for controller work, often including closing the books, resolving accounting issues and advising on the numbers. An AI controller is software that does a controller’s daily work across a firm’s client book: monitoring cash, keeping forecasts current, testing scenarios, checking budgets, explaining changes and tracking what was agreed. The person brings judgement and accounting depth. The software performs the daily work for every client without its hours growing as clients are added.

Can an AI controller do the job of a fractional controller?

It can perform the daily controller work that often takes up much of a fractional controller’s time. It doesn’t keep the books, resolve accounting issues or own recommendations. Whether that’s enough depends on whether your firm is short on daily controller work, on judgement, or on accounting capacity.

Does a CPA firm need a fractional controller if it uses an AI controller?

Not necessarily. If the accounting is covered and the firm’s CPAs or CFOs already own financial review and recommendations, the gap is often the time daily controller work consumes, and an AI controller performs that work. If the firm still has more review, accounting or client work than its people can handle after that, another person may still be needed.

When should a firm hire a fractional controller?

When nobody in the firm owns ongoing financial review and recommendations for clients, when closing the books or resolving complex accounting issues is the gap, or when human work still exceeds capacity after the daily controller work is automated. Pairing that person with an AI controller means their hours go to judgement and accounting depth rather than daily controller work.

Does an AI controller keep the books?

No. It works from the records bookkeeping produces and assumes bookkeeping, reconciliation and closing the books are covered. It can detect problems in those records, such as a mismatch with the bank. Resolving them is accounting work for a person.

I own a business. Should I hire a fractional controller or use an AI controller?

AgentLink’s AI controller is offered through CPA and fractional CFO firms rather than sold directly to owners, so the practical question is whether your firm offers it. If no firm provides ongoing financial review and recommendations for your business, a fractional controller or fractional CFO is the place to start.

About the author

Nilanjan Raychaudhuri is the founder of AgentLink (Tublian LLC, Columbus, Ohio), which builds an AI controller sold through CPA and fractional CFO firms. He has spent the past year working with practitioners at CPA, accounting and fractional CFO practices on how controller work is actually delivered between calls. Team page

The practitioner observations in this guide come from design-partner conversations. Corrections to nilanjan@agentlink.finance.