Where Does The Time Go?

Reading Time: 8 minutes

Most people I know are paid for time. It’s an artificial contract. Notionally, I am paid for 40 hours of work a week. There is some expectation that those 40 hours will occur within a specific schedule, like 9am to 5pm. Employers pay me for my time because they have neither the ability nor interest to try to calculate whatever my value might be and because time is how our society approximates work output. As I have become more expert, I have needed less time to accomplish some of the same things that I achieved when I was newer to my work. I have also found ways to change my work, either by automating it or by eliminating unnecessary make-work. If we consider the place that artificial intelligence has in the workplace—assuming it does not eliminate a role but is used within one—we should also be thinking about who gets the benefit of the time savings.

People talk about the end of the billable hour for the legal profession but I don’t see it. It’s not a lawyer-specific method of measuring value. Most people are paid by the hours that they spend at work. Freelancers are often paid for their time spent, and people, like doctors, who may not be paid for their hours are still paid based on the number of activities they can accomplish in a set time period.

Lawyers may be notorious for abusing the value of the hour—in the case of the $3,000-an-hour lawyer, fools and their money are soon parted I guess—but most people could calculate what their current hourly pay is, lawyer or not. The billable hour is just how a lawyer creates an artifact to persuade a client who is not getting a tangible widget that intangible widgets were created. Not all clients are persuaded.

They may not approach it the same way but, once a doctor knows that an office visit is 30 to 39 minutes, they can price it even if they can’t measure the value of that time. The same thing goes for lawyers who have alternative fee structures: those are based on time spent to accomplish tasks, priced out to a flat fee. Lawyers who bill 2,400 hours a year—10 billable hours a day—are generating a certain amount of revenue, whether the hours are discrete or not. It is certainly not based on quality or there would only ever be happy clients. Whether someone has a time card that captures their presence down to the minute or whether they are considered salaried, we work to a clock.

In a commercial setting, the goal is always to move more widgets, whether those are wristwatches or motions to dismiss. Any time saving is immediately converted into more widgets. Lawyers might argue that they’re being paid to think but that’s just a load of old tosh. Many people without the billable hour think about work outside of their designated work time. There is no incentive to save time other than to make more money, either for the worker or, more likely, for the organization. We have hardly left Frederick Taylor behind although, in law firms, the lawyer’s value to the firm is based on time spent but the individual lawyer values themselves by their wealth creation or at least not for abstractions like “justice”.

I’ve been in salaried roles my entire law library career. I understand the perspective that a salaried role is, in essence, ungoverned by time for positions that are “exempt“. You work your notional 40-hour week but if there is more work to be done, you do it without additional compensation. In essence, there are some people, both employers and employees, who view salaried jobs as an unlimited amount of time (the time the job “takes”).

That’s where a manager comes in.

Time Lapse

First, any leader who assumes that their staff is going to work at 100% productivity, 100% of the time is fooling themselves. The vision in my head of this sort of workplace is this scene from Schindler’s List:

We can ask our people to work at high levels of productivity, even exceptional levels. But not all the time. It doesn’t even require a system to break or need to repair to cause lower levels of productivity. The manager’s role is to watch priorities and productivity and help staff to be successful. The success should be measured not in completing a certain number of hours but in creating value and completing objectives that contribute to the organization’s mission. The latter, what is achievable, should be tailored to fit into the former if we require people to work to a schedule.

Every job has the potential to create 24-hours-a-day-worth of widgets; the 40-hour work week is just a modern cap on that. There is always more work to be done, whether it is handling what is currently known or developing new skills or services for the future or for expansion. This can be particularly true in workplaces like law firms or law schools where there are no perfectable outcomes but people are trained to think there might be one. Where people will perform legal research or drafting until it has become a pointless, repetitive (and yet billable) exercise, where satisficing may not be valued or modeled.

A manager needs to make sure that people have the time to create the widgets that are necessary, to provide the services that are a priority. They need to ensure people are able to reduce the tautness in their line, whether that is just staring out a window or reading something that helps them get better in their own role or expertise.

One method that has always stuck with me was the “side project” approach, although it appears to have fallen out of favor. If we know that the amount of optimal time our people can work is less than 40 hours (100%), then perhaps we build that in to the expectations. Assume the amount is 80-85% where people are on top of their primary roles and let them use the other 15-20% to expand their personal expertise and work on things that are rejuvenating in ways that their primary role may not be.

20% is 8 hours. I believe that if you have an expectation that people will be giving 100% of their energy and focus for 100% of their hours, you are likely to find that they will build in at least 20% slack regardless. There is no incentive to do otherwise. They know, even if the manager doesn’t, that running taut in all weather isn’t a sound strategy. Sometimes you need to luff.

Time Saving

Enter automation. At the moment it is generative AI and agentic AI, two still emerging technologies that promise time savings. We have had the ability to automate for decades and the biggest difference is that we may be able to delegate work tasks to these tools without anticipating every step of the resulting activity. That delegation is still bounded, constrained. Lawyers are subject to professional rules, law schools are bound by privacy laws and other regulations, and all law librarians work within boundaries like unauthorized practice of law and avoidance of giving legal advice.

So long as a human needs to remain in the process, automation tools can only deliver up to 100% of the same output, as measured in hours. That percentage will vary depending on so many different criteria that it is impossible to pin down. What are the risks? What are the costs? If I use artificial intelligence and delegate work to it, what are the impacts to my processes? What new processes do I need to insert to ensure the delegated work is correct?

This is an interesting post about reducing human oversight in the AI process, suggesting that there is a “scalability trap.” This is trickier in a field like law where every output needs to be verified by the lawyer. What happens when a lawyer delegates the physical filing of a document to a staff person and the person fails to file it? It’s an obviously clerical task; I did that for a couple of years as a teenager for a law firm. And yet the lawyer is responsible for making sure it occurs and a delegated failure involving a human may become indefensible if it’s delegated to an AI who skips an e-filing deadline.

Let’s focus on the time, then. We have an individual contributor who is expected to deliver 40-hours worth of work. They adopt generative AI for drafting a document and agentic AI for managing their personal calendar and for performing email triage. The savings on a weekly basis in time is 4 hours for drafting and 2 hours for scheduling and email management. At least at the start, some of that time savings is then lost as the drafting needs to be verified, the calendaring and email triage needs to be validated to ensure appointments are made correctly, email is deleted or organized appropriately. Let’s assume that, over time, these delegated functions work without oversight, just as they would if they had been delegated to a human who gained expertise in them.

What does that person now do with those 6 hours of saved time?

No one is measuring value generated or we wouldn’t even be talking about time. This is one reason I have no problem with the employees who work two full-time jobs. For whatever reason, their employers are so fixated on time that they do not realize that the workload can be done in far less than 40-hours, that the person is able to perform multiple roles within the time expectation. I’m not really fine with that—it seems unhealthy at a minimum and I would worry about focus—but it’s a management problem, not an employee problem.

But what if it isn’t a problem at all. What if an employee’s role can be accomplished with the help of automation in 34 hours instead of 40? Or 20 hours? The default expectation is, I think, that the employer should get the benefit of that automation: 6 more hours of widgets. Never mind that automation is an amalgam; it requires the tools, for sure, but it still requires someone with the necessary expertise to ensure the output is compliant.

Since the employer is not really measuring quality but quantity—widgets moved, hours worked even if we call that “hours billed”—why shouldn’t the employee get the benefit? If a lawyer can accomplish 40 hours of “work” (however that is defined) in 32 hours, why wouldn’t the lawyer be entitled to the financial value of that 40 hours? If you’re a solo lawyer and you are meeting financial goals (salary + overhead + growth), and you can do that in 32 hours, it would seem as though you could stop at 4 days a week. I get that large law firms are designed to extract maximum hours from people in lieu of value, so there is no incentive for a big firm lawyer to be more efficient, either for the client or for the law firm, since either one or the other will not reward that efficiency.

A lot of this post grew out of reading a story that discussed John Maynard Keynes expecting we’d be working 15-hour weeks (this NPR story gets at it more directly). But once we start valuing our time at a particular dollar amount (instead of the work we do having its own value), perhaps we see time as “wasted” as a financial loss. And it is, in a sense, if we save ourselves time but our employer pays us less because of that fact. If we are using automation tools to perform the tasks for us, delegated with oversight, then the saved time should be compensable even if we are not doing it ourselves.

The American Bar Association’s formal opinion on artificial intelligence suggests the lawyer cannot bill for the saved time (emphasis added):

This seems a bit toothless in a profession where a lawyer’s $3,000-an-hour fee is considered “reasonable”. But I digress.

I think this is a terrible outcome. It makes the assumption that “little or no work” has been done merely because the automation has completed some steps of the work. Expertise goes into making processes happen. If I am a slow typist and can draft an email with automated tools faster (whether it’s AI or the use of macros, document automation, and templates), why shouldn’t I be able to charge for that email the same as I would have if I’d typed it manually? The same thinking and expertise is going into the final product.

I expect that we will see pricing change to adapt to that. The output is the email. If it costs $100 for an email right now, without automation, I am sure it can be re-priced to $100 with automation; the human components will just be priced higher.

It begs the point of automation, though. It’s not clear to me what the incentives are in law practice. You can’t be compensated for being automated, other than to take on more un-automated (and thus billable) work. In other words, you are still going to have to work 40-hour weeks even if you can accomplish a 40-hour output in 32 hours. You can bill that additional six hours—and those on the 2,400-hour-per-year are unlikely to see that change because the point is extraction not value—but to what end? Wealth creation continues to balance against the hazards and harms the legal profession inflicts on its practitioners.

This is likely to be the outcome in other workplaces. A law librarian who figures out how to use low-code and automation tools to reduce administrivia isn’t going to be incentivized for that. They will continue to be valued by the number of hours they are visible and available, whether or not they are using that time in a way that is valuable, either for the employee or the employer. I would expect to see automation used to expand the space between the 100% time and the 80%, whether or not that is visible to the organization. It may be an opportunity for people to make time for themselves.