By Michael Domanic, Section Head of AI

If AI saves your employees 10 hours a week, what happens to those hours?

Do you expect people to produce more of the same work? Improve the quality of what they’re already doing? Take on new work they couldn’t do before? Or do you use the productivity gains to cut contractors or FTE headcount?

Most companies are seeing some individual productivity gains from AI. In Glean’s Work AI Index 2026, respondents reported saving about 11 hours a week with AI. At Section, our most proficient users are creating enough capacity to double, and sometimes triple, their output.

Far fewer companies have figured out what the extra time is for. It’s usually left up to the individual. And without a strategy, the time goes to output – aka creating more stuff.

If you don’t decide where the time goes, you’ll get more stuff

Judgment used to be free, because making things was expensive. If it took two weeks to write a report, you’d usually ask, “What’s the point of this report?” first. Now that you can build it in an afternoon, it seems like a no-brainer (regardless of the actual value). 

As people get better at AI, they start producing more of everything. More dashboards, analyses, presentations, agents, mockups, etc. Some of it is valuable - but a lot of it exists just because someone can make it quickly.

Take this newsletter as an example. If I started it from scratch every week, it would take me about 6 hours to write. With AI as a thought partner I’ve gotten that down to three. With the time I’ve saved, I could write two newsletters a week - so should I?

Maybe - maybe not. It depends on the value the newsletter drives, the incremental increase in value from doubling that output, the trade-off of what else could be done with those three hours, etc. It might be the right choice, but it shouldn’t be the default.

How to decide where to put the time

Every hour AI saves can go one of a few places.

To pick the right one, you need to know what’s limiting your results - which means you need to know the results you’re trying to drive. For a content marketer, this could be lead volume, website traffic, meetings from SEO, etc. For a salesperson, it could be conversion to closed-won, ACV, or time to close. 

Once you know the metric you’re trying to drive, ask yourself: What problem is stopping us from improving it? Here are three common ones.

The problem: Capacity. You know what works, and there’s demand for more of it. You just don’t have enough time or people to do it.

How AI can help: Do more of the same work. If your customer support team has a backlog of tickets and consistently high resolution rates, AI can help agents resolve more issues without sacrificing quality. The key is knowing that additional volume will actually improve the business outcome.

The problem: Quality. You’re producing enough work and you have evidence (from competitors, etc.) that it should work, but it isn’t delivering the results you need. More of the same won’t help.

How AI can help: Do the existing work better. If your engineering team is shipping features quickly but releasing too many bugs, AI can free up time for better testing and code review. The goal isn’t to ship more features. It’s to improve reliability and reduce the cost of fixing problems later.

The problem: Untapped opportunity. There’s something valuable your team could be doing, but it hasn’t been possible with the time or resources available.

How AI can help: Take on new, higher-value work. Maybe your finance team has never had the capacity to build detailed profitability analyses by customer or product line. AI could make that work feasible. But you still need a business case: What decisions will the analysis inform, what financial impact could those decisions have, and is that worth more than other uses of the time?

There’s one more question I’d ask before choosing any of these options: Is the work we’re already doing worth doing?

One of our account managers recently used AI to analyze outcomes across our legacy customers. She discovered that one of our standard activation events wasn’t producing the results we expected. In fact, customers who skipped it were performing better.

She could have used AI to make the event easier to deliver (or do more of them!). Instead she determined it wasn’t worth doing at all. Sometimes the right answer is simply to stop doing the work and capture the savings. Not every hour AI frees up needs to be reinvested - it can be cut via headcount freezes or cuts (first to agencies and vendors, not FTEs). 

Build a scrappy process for deciding

Most decisions about how to use the time AI gives back will be made by individual employees and their managers. You can’t wait for a biannual planning process to make them. You need a simple habit that fits into regular 1:1s and team meetings.

I’d start with a two-minute value check. Before someone decides where to put their newly available time, they should be able to answer three questions:

1. Who benefits from this work?
2. What should it change? 
3. In the next 30 days, how will I know whether it worked? 

If they can’t answer those questions, they probably need to think harder about whether the work is worth doing. If they can, write down the answers and come back to them in a month.

Managers should reinforce this thinking in everyday conversations. When someone shows you an agent they built, ask who’s using it and what they expect it to change. 

And once a month, spend 15 minutes as a team reviewing what happened. Where did the time go? What created value? What didn’t? What should you stop doing?

The goal is to help employees think in context of the business metrics you’re trying to drive - and develop better judgment about where they spend their time.

Have a great week,

Michael

Your fellow Head of AI

‍