LONDON — On April 23, I started work at 8:49 a.m., reading and responding to emails, browsing the news and scrolling Twitter. At 9:14 a.m., I made changes to an upcoming story and read through interview notes. By 10:09 a.m., work momentum lost, I read about the Irish village where Matt Damon was living out the quarantine.

All of these details — from the websites I visited to my GPS coordinates — were available for my boss to review.

Here’s why: With millions of us working from home in the coronavirus pandemic, companies are hunting for ways to ensure that we are doing what we are supposed to. Demand has surged for software that can monitor employees, with programs tracking the words we type, snapping pictures with our computer cameras and giving our managers rankings of who is spending too much time on Facebook and not enough on Excel.

The technology raises thorny privacy questions about where employers draw the line between maintaining productivity from a homebound work force and creepy surveillance. To try to answer them, I turned the spylike software on myself.

Last month, I downloaded employee-monitoring software made by Hubstaff, an Indianapolis company. Every few minutes, it snapped a screenshot of the websites I browsed, the documents I was writing and the social media sites I visited. From my phone, it mapped where I went, including a two-hour bike ride that I took around Battersea Park with my kids in the middle of one workday. (Whoops.)

To complete the experiment, I gave my editor, Pui-Wing Tam, the keys to the Hubstaff program so she could track me. After three weeks of digital monitoring, the future of work surveillance seemed to both of us to be overly intrusive. As she put it, “Ick.”

Week 1

Adam: I downloaded Hubstaff to my laptop and phone with more than a touch of skepticism. I had heard of this type of tool being used by Wall Street firms for years, mainly in the name of security, with employees rarely having any say about how they were being watched.

Dave Nevogt, a founder and the chief executive of Hubstaff, who gave me a free trial to test its subscription software, said work-from-home orders in the coronavirus outbreak had made employee-monitoring software a hot ticket. Trials of Hubstaff software, which cost $7 to $20 a month per user, have tripled since March, he said.

“The world is changing,” Mr. Nevogt told me. Workers know they are being watched, so it does not violate privacy, he added.

One main feature of Hubstaff is an activity monitor that gives managers a snapshot of what an employee is doing. Broken down in 10-minute increments, the system tallies what percentage of time the worker has been typing or moving the computer mouse. That percentage acts as a productivity score.

I tried to embrace the feedback. Each day, an email was sent to me and Pui-Wing with an overview of my day: hours worked, the productivity score, and the websites and apps that I was using.

One day last month, when I was putting the finishing touches on an article, I spent 3 hours and 35 minutes editing the document, and an hour inside a file holding background research and interview notes. Another 90 minutes were spent on email.

This was one of my more productive days, but the software still tallied my deviations. It showed I was on Twitter for 35 minutes and lost 11 minutes browsing Spotify. Slack, the collaboration tool, swallowed 22 minutes. Other days, food was a common distraction, including one 10-minute hunt for takeout pizza.

Hubstaff also logged my GPS coordinates, a feature that Mr. Nevogt said was mainly used by companies trying to ensure that their salespeople were visiting clients. Given that London has been on lockdown since late March, I had few movements to track. The software mainly caught me jogging around a nearby park. And going to a wine shop.

Week 2

Adam: Once accustomed to life under surveillance, I made the questionable decision of letting Pui-Wing have access.

“You are agreeing not to fire, judge or blackmail me for whatever this turns up,” I wrote to her in an email beforehand.

Pui-Wing: I was curious, I admit it. But also reluctant because do we really want to see someone’s minute-by-minute location or how often he or she uses Twitter?

With those misgivings, I opened the program and saw a dashboard. It showed various categories, including screenshots of Adam’s computer, his time sheets, apps and URLs he had visited and his whereabouts.

I clicked on screenshots and saw that Adam had been online for 9 hours 42 minutes 17 seconds the previous day. The dozens of screenshots included those of a Google Meet conference call that Adam had participated in, which displayed as extremely close-up photos of the faces of numerous colleagues.

I quickly retreated to the main dashboard. There I saw that Adam’s activity for the week was at a somewhat disappointing 45 percent. He later explained that the number didn’t accurately reflect his time spent working because it logged only when he was typing, not when he was making phone calls or doing other work away from his computer. Right.

Adam: For employers nervous about wasting money in a shaky economy, I could see Hubstaff’s appeal. Mr. Nevogt introduced me to Chris Heuwetter, who runs a marketing company in Jupiter, Fla., called 98 Buck Social.

Mr. Heuwetter said he had seen work hours collapse after he let his 20 employees work from home in the virus outbreak. The company was facing a drop in sales, but Mr. Heuwetter said some employees did not start responding to messages until after 10 a.m. Responses to customer questions also slowed.

So he began using Hubstaff on March 31. Once he did, he said, his employees’ productivity levels rose “immediately.”

I could relate. Hubstaff was starting to affect my behavior. Each day, I logged in early because it was keeping a running clock of my activity. Knowing my online actions could be reviewed, I did not spend (as much) time reading about sports and rarely opened messaging apps on my laptop, nervous about a screenshot catching a private exchange.

But my activity scores stayed stubbornly low, usually from 30 percent to 45 percent. On April 14, Hubstaff showed that I worked for nearly 14 hours but had a productivity score of 22 percent.

Week 3

Adam: The moment when I no longer wanted to be monitored came on April 23 at 11:30 a.m., when Hubstaff caught me doing an internet exercise class. By the time I realized I had not logged out, it had snapped a screenshot of the trainer setting up to teach the class in her living room.

Even though this was just an experiment, it didn’t make it any less embarrassing and intrusive. And it goes beyond being caught exercising in the middle of the day. What if other screenshots exposed sensitive health or financial information?

I trust Pui-Wing, but the monitoring systems have few safeguards to prevent abuse, and they rely on managers exercising judgment and restraint.

Pui-Wing: Fortunately, I did not see Adam’s internet exercise class. After poking around the Hubstaff metrics, it was clear it did not capture when he was reporting and talking to sources. It was thus irrelevant — at least to how we work.

Also, did I mention it was yucky to see so much of someone’s information? I didn’t log back in.

Occasionally, I glanced at the daily emails that Hubstaff sent about Adam. They showed his productivity score at 30 percent, sometimes edging up to 50 percent. I chuckled when I noticed that he began spending more time on news websites as his behavior changed.

Adam: By the end, I found myself trying to cheat the Hubstaff system altogether. As I write this at 11:38 a.m. on April 24, I am about to get some coffee and spend time with my cooped-up kids. But I plan to leave a Google Doc open on my computer that Hubstaff can screenshot to make it look like I was doing work.

Even if my editor says she isn’t looking. Just, you know, in case.

Pui-Wing Tam contributed reporting (and surveilling) from San Francisco.