Will AI Break Your Labor Efficiency Targets?

✋Welcome to The CAS Cache, a newsletter designed to help accounting firms grow their CAS offerings in five minutes or less.

Disclaimer: Some links below support my writing of this newsletter, and some give you a deal.

Issue Sponsor 😎

Don’t outsource. Hire direct in the Philippines.

You can hire an accountant who works for you and only you. No middlemen. No ongoing fees. Just you and your team. Schedule a call with TeamUp now to learn how direct hire works.

Want to sponsor The CAS Cache? Reply to this email to get more details.

A colleague recently reached out after seeing that I will be presenting on Using LER to Forecast Growth, Staffing, and Profitability at Intuit Connect in October. She is rebuilding her firm’s financial model and asked a question I think a lot of accounting firm leaders will be asking soon:

As AI changes how we deliver work, will the benchmarks we use to measure healthy labor efficiency and margins need to change too?

My honest answer is probably, but I do not think we are there yet.

For now, I would keep the targets the same, track results monthly, and reassess them quarterly. AI adoption is moving quickly, but the financial impact will not show up at the same time or in the same way for every firm.

That is exactly why LER and other metrics are worth watching to compare company historical trends rather than industry benchmarks.

A quick refresher on LER

I have written before about using the Labor Efficiency Ratio (“LER”) to move advisory conversations beyond hours. At a high level, LER measures how efficiently a business utilizes labor. I typically look at three calculations with CAS clients:

Direct Labor Efficiency (“dLER”): measures efficiency at the service-delivery level. I have found $3.00 or better to be a useful starting target for many professional service firms, including accounting firms.

Management Labor Efficiency (“mLER”): measures management labor efficiency and can help identify when leadership becomes too heavy relative to production. This is usually the biggest problem I see for small businesses as they grow. I have generally found $3.50 or more to be a useful starting point in professional service firms, such as accounting firms.

Overall LER combines the full labor picture. I use it with every client, especially smaller businesses, and generally look for $2.00 or better.

Those are not universal rules. I start with the client’s historical performance, review how contribution margin and net operating income behave, and then establish targets that fit the business.

From there, LER becomes useful for answering questions like whether the business can afford another hire, how much revenue growth is needed, or what cost reductions would be required to reach a target profit.

You can read the full breakdown in Turning Labor Efficiency into Advisory Gold.

Will AI change the benchmark?

Eventually, I think it will. The bigger question is when.

I have not been factoring AI adoption directly into the LER formula because most firms are still in the investment stage. They are buying tools, testing workflows, training teams, and experimenting with new ways to deliver work.

The efficiency may be starting to show up, but that does not necessarily mean the labor savings have reached the financial statements.

I have one client who hired a full-time AI employee whose sole job is to work on various AI projects. We do not track LER for this particular client because the company has historically overpaid and overhired staff, making the metric less meaningful without additional normalization.

But if we did calculate it today, I am confident LER would initially decline. The business added another salary before eliminating any labor through efficiency gains. That means labor cost increased while revenue and staffing requirements have not yet had time to adjust.

This is an important point for firms investing heavily in AI. Your efficiency metrics may worsen before they improve.

That does not automatically mean the investment is failing. It may mean the firm is carrying the costs of its old and future delivery models simultaneously.

The real test comes later.

Are AI projects reducing the time required to serve clients?

Is that capacity being converted into more clients, higher-value work, or fewer labor needs?

Is revenue growing faster than labor cost?

Are margins improving?

Until those outcomes start appearing consistently, I would not lower or raise the LER targets simply because a firm says it is adopting AI.

Keep the benchmark steady. Let the monthly trend give you the first signal. Then use quarterly reviews to decide whether the target still reflects a healthy operating model.

Where should training and internal meetings go?

My colleague also asked whether team training and internal meetings focused on client delivery should be included in cost of service or overhead.

I could argue this both ways, but my preference is to include the full salary of someone in a producing role in cost of service.

Training is intended to improve that person’s ability to deliver the service. Internal meetings are time they could otherwise spend performing client work. From an economic perspective, both are part of the cost of maintaining delivery capacity.

I also hate meetings, so less is more in my book. Including that time in the cost of service may decrease the margin, but that is not necessarily a bad thing. It shows the real cost of the delivery model and creates an incentive to question whether every meeting is necessary.

The goal should not be to move costs into overhead so the service margin looks better. The goal should be to understand what it actually costs to deliver the work.

At the same time, training is a real and necessary investment. That cost should be considered when setting prices.

In my article on Creating a Floor Price, I discussed building development time, paid time off, fair wages, capacity, and target margins into the minimum engagement price. The idea is to ensure your pricing supports quality delivery without requiring unrealistic utilization of the team.

If training is necessary to serve clients well, client pricing must support it.

What about bonuses and incentives?

I would track bonuses separately from normal labor cost. Base compensation reflects the recurring cost of delivery. Bonuses are variable and should usually depend on a broader set of outcomes.

I am a fan of subjective bonus plans that include objective guideposts.

LER can be one of those guideposts, but I would not make it the only one. I would also want the company to achieve more than 10% profit before meaningful bonuses are paid.

That balance matters because pushing one efficiency metric too aggressively can create bad behavior. Teams may avoid necessary training, cut corners on service, or resist helping colleagues because they are focused on protecting their individual numbers.

LER should inform the bonus discussion, not control the entire plan.

A good bonus plan should consider efficiency, company profitability, client experience, teamwork, and the employee’s overall contribution. Otherwise, you risk turning LER into another version of the billable-hour monster.

The takeaway

I expect AI to eventually change healthy labor and margin benchmarks. I just do not think CAS and accounting firms should guess at the change before it appears in their numbers.

Keep your current targets for now. Track LER monthly. Review the targets quarterly. Separate the cost of AI experimentation from proven efficiency gains, and watch whether added capacity turns into revenue growth, stronger margins, or lower staffing needs.

Thanks for reading, Luke Templin!