AI Is Changing Firm Compensation. Now What?

I run customer operations at Pilot, which means I manage the managers who manage the managers who manage the accountants who close our clients' books. I hired most of that team. I trained them, I built the career ladders they move through, and I sit with our managers every week on how we run the firm and where it goes next.

So when I talk about AI in accounting, I'm not talking about it as a technologist. I'm talking about it as the person who has to get a few hundred accountants to actually use the thing. And the lesson I keep relearning is that the software is the easy part.

The thing most firms forget when implementing AI

Most firms implement AI and forget about the people who use it. More specifically, they forget to ask whether those people have any reason to.

If your team is paid and measured the same way they were before the software arrived, you have changed nothing about how anyone behaves. You have a new tool sitting inside an old incentive structure, and the incentive structure always wins.

David Maister’s Managing the Professional Services Firm, says every firm exists to do three things: deliver an excellent client experience, give staff a career that progresses, and generate enough profit to fund the partners' lives. Most AI strategies focus on either the client or profit pieces, whereas staff training on AI becomes a webinar.

Your employees are not the audience for your AI strategy; they are essential to it. They will end up either the technology's biggest advocates inside your firm or its biggest opponents, and which one is usually decided before the tool goes live.

Here are four people-centric portions of your AI strategy to consider:

1. Stop paying people for hours

Your clients have AI subscriptions too. They know what an AI agent can do in a few minutes, and they aren't going to keep paying for the hours it used to take a person to do the same work by hand. Every client you serve will expect you to work as efficiently as the technology allows, and they'll be right to.

That makes the billable hour very hard to defend. Clients were never really buying your effort. They were buying the outcome: a closed month, financials they can trust, an answer to the question they asked. Firms should charge for the deliverable, because that's what they're paying for.

Once you price that way, hours stop making sense as the basis for pay too. If your accountants are measured on hours, tasks completed, or utilization, every hour the software saves them is an hour they're penalized for. Pay people for the outcomes you care about most, which for most firms are revenue and retention:

• Revenue: Give accountants an earnings tied to the revenue of the portfolio they manage. When automation lets someone serve more clients well, or a client grows into more services, their book grows and so does their pay.
• Retention: Tie a real share of pay to whether clients stay. It's the most honest measure of service quality you have. A client can give you a good survey score and still leave, but a client who renews year after year is telling you something you can trust.

Together, these point your team at the same thing the technology is for: serving more clients, and serving them well enough that they stay.

2. Train on AI as its own skill

Your people know accounting. Most of them don't yet know how to get a good outcome out of an AI tool. Often an accountant will try the tool, give it a vague instruction, get a mediocre result, decide it doesn't work, and instead goes back to doing all their work by hand. Now you've paid for the software and paid for the manual work.

The good news is that’s a training gap that can be fixed. Here are three things to teach:

• How to prompt: Your accountants need to know what context to give the AI they’re using, how to ask for a specific output, and how to check the answer. Most people pick it up in a few sessions with prompt training.
• Which tool for which occasion: A general-purpose model, a purpose-built accounting platform, and a spreadsheet are three different instruments. Your team needs to know which one to use for specific occasions.
• How to check the work: Working with AI productively means knowing what a correct result looks like before you ask for it, so you can tell when you didn't get one.

One thing to also factor in is the type of tool you choose. It’s important to pick tools that are easy to adopt. Many AI tools in the accounting space require you to build skills or have deep technical expertise to use. The lower the learning curve, the better adoption.

3. The client is the reason

Most AI rollouts introduce the tool to the team as an efficiency play: here's the software, here's how much time it saves. That's a hard pitch for the team to get excited about. For an accountant worried about their job, "saves time" can sound a lot like "needs fewer of you."

Technology is worth the investment when it serves a real market need, which means it lets you give clients more of what they want. Tell your team that story, because it's the one that gives them a reason to use the tool.

Your accountants already know what clients want. They're the ones fielding the question that's gone unanswered for a week, apologizing for a close that ran late, or wishing they'd had time to flag a cash problem sooner. Ask them where they'd put the time back. They'll point the technology at what clients value more accurately than any rollout plan will.

Then connect the tool to those outcomes in plain terms. Not "this automates reconciliations," but "this gets your client's close done days sooner" or "this frees you up to call them before they call you." People adopt a tool more willingly when they can see what it does for someone they care about, and most accountants care a lot about their clients. It also lines up with how you're paying them now, because what clients value is what drives revenue and retention.

4. Build business judgment in every seat

When automation takes over production work, the obvious move is to point everyone at advisory. For some of your team, that's the right call. But not every accountant wants to spend most of their time sitting across from a client, and the time you free up is worth more when it goes where each person is strongest.

What every accountant needs is a better understanding of how a business actually works. When I think about training my team now, it's less about how to write a journal entry to correct a past error and more about the full picture. They need a good intuition for what a clean set of financials looks like versus a bad one. That means teaching people to read the reports they produce and what those reports say about a client's business, to spot a trend before the client does, and to notice when the cash position needs attention.

For your client-facing people, that judgment becomes advisory: explaining what the numbers mean and what to do about them. For the people running a large book, it becomes the ability to tell when something the system produced doesn't fit the business behind it. Clients pay for both.

These skills take time, expertise, and taste, and nobody develops them by accident. When you invest in them, your team gets a career instead of a job, and your firm earns more because every client gets more value.

The software is the easy part

Every firm can buy the tools right now. What decides if they work is whether the people using them have a reason to, which is a combination of their pay, their skills, and their next role.

Your team will either be AI’s biggest fan or its biggest adversary, so you need to give them a reason to use it.

Sponsored Content: This article is generously brought to you by one of our valued sponsors. Their support enables us to continue delivering expert insights and the latest industry trends to our dedicated community of accounting professionals.