Concept

Automation can take away one part of a task while leaving work around it, or adding some.

Issue

A quick result can hide the time spent getting ready, checking and keeping the process working.

Solutions

Understand the work and the benefits involved, then make a complete time account.

Say you’ve automated a routine report. The draft appears quickly, but you still get the inputs ready, check the result and fix the odd failure. You want to know whether the whole job has got easier.

To answer that, look past how quickly the draft appears. Count the work you no longer do, the work that’s left and the new work you’ve taken on. A faster step and a shorter working day are different results.

This module is for working adults who want to account for the ongoing work that comes with routine automation, at home or in the office.

What you’ll get from this module

In this module

By the end of this module, you’ll know what to count when you compare work before and after automation. We’ll link a study of robot vacuums at home, a paper on automation in industry, research on smart homes and a writing experiment. You’ll choose a situation from daily life and work out the time involved, the benefits and any costs you still need to understand.

What happens to the work around automation?

Household field study: machine time and human time can move in different directions.

Emilie Vrain, Jan Bieser and Charlie Wilson’s 2026 study followed ten chosen households in and around Oxford, UK. The households tried a robot vacuum for 30 days. They were measured before the trial, and followed up 15–18 months later. They weren’t put into automation and comparison groups by chance.

During the trial, the time people spent cleaning by hand fell by 45%, an average drop of 24 minutes a week. The total time spent cleaning floors rose by 189%, but that total included the time the vacuum was running. It wasn’t a measure of how long people worked.

The difference matters. A machine can spend longer cleaning while you spend less time cleaning yourself. Counting all of that machine time as human work would wipe out the saving in hand cleaning that the study recorded.

Yet the chart of cleaning time left out getting ready, keeping an eye on the vacuum and looking after it. People described work such as tidying up before a run and rescuing the vacuum when it got stuck. The saving in hand cleaning was real within what was recorded. But the full balance of human time was still incomplete.

Why it matters: The study also kept using the vacuum apart from using its app. Two households that used the vacuum later stopped using the app and pressed the button on the vacuum instead. Optional work controlling a device and the task itself can be different parts of an automation.

Read the household study of automation and time use.

Conceptual paper on people and automation: doing the task is only part of the job.

Lisanne Bainbridge’s 1983 paper looked at control of industrial processes, with examples from automation on aircraft flight decks. She described two roles that can be left for the person in charge: watching the automatic system, and stepping in when it isn’t working properly.

Bainbridge explains a problem with automation in industry. She isn’t reporting an experiment that timed tasks. Her distinction is useful: doing a task and checking that it worked are different jobs. Automating the first can leave someone responsible for spotting problems and sorting out a response.

That helps explain why a finished output isn’t the whole workload. Someone may still need to check that the result can be used. And when the system can’t handle something, someone has to deal with the work it left unfinished.

Read the paper on the ironies of automation90046-8).

Research on smart-home experiences: a useful tool can still need upkeep.

Yolande Strengers, Jenny Kennedy, Larissa Nicholls and Paula Arcari’s 2018 report looked at 31 Australian smart-home households who chose to take part. The researchers interviewed 42 people, toured their homes and watched how they lived. These early users weren’t typical of households in general.

The report describes the work needed to keep the technology running, up to date and working together. It also describes convenience, enjoyment and support for people with disabilities. Most households said they didn’t think their technology saved them time or money overall. That was how they saw it, not a measured balance of time.

Why it matters: Those experiences show why time alone doesn’t tell you whether a tool is worth having. It can need looking after and still help someone do things more comfortably, or more on their own.

Read the report on Australian smart-home experiences.

What does automation leave for you to do?

Randomized writing experiment: substantial task savings are possible.

Extra work doesn’t mean automation has to lose the comparison. Shakked Noy and Whitney Zhang’s 2023 online experiment gave writing tasks to 453 college-educated professionals, each task tied to their kind of job. Half were picked by chance to use ChatGPT.

Their published summary reports that, on average, the tasks took 40% less time and the quality of the work rose by 18%. These tasks didn’t need exact facts or knowledge of a particular workplace. The result is about those writing tasks. It isn’t about the ongoing upkeep of an automated way of working.

Why it matters: So AI help can save time and improve quality on the same task.

Read the writing experiment’s published abstract. Read the authors’ manuscript for the study methods and limits.

Give a fair answer about time saved

What this means for you

Start with the result you need, then count the work people do to get it. Write down the benefits too. Saving time is one reason to use automation, but it isn’t the only one.

Compare the same result over a stated period

Describe the output, how much of it you need and what counts as good enough. Then choose the period your account covers, and write down how many times you expect to use the automation in it.

Say your automated report now has more detail than the old one. Write down the work that fuller report takes. Then, separately, compare the work needed for the same amount of detail as before. Otherwise, you might mistake making more for failing to save work.

Count the getting ready, checking, fixing and coordinating you did before automation. Checking you already did belongs in both accounts if it still happens. It doesn’t become new work just because automation makes you notice it.

Follow the work before, during and after execution

Keep setting up apart from the work you keep doing. Setting up means getting the system ready, learning to use it and testing that it works. Say how much of that one-off time you count in the period you chose. If you spread it across the times you expect to use it, say so clearly. Don’t leave setting up out, and don’t count its full cost again every time.

Then list getting ready, keeping an eye on things, checks, fixes and problems the system can’t handle. Include upkeep, updates, broken connections and the work of moving to a replacement system, where they apply. For each activity, name who does it, how often it happens and how much time it takes.

For work you repeat, multiply the number of times it happens by the time each one takes. That gives its total for the period. Count each activity once: a fix you’ve already included under checking shouldn’t turn up again under repairs.

Write down the steps you still do by hand, and any work that’s moved to someone else. If you save time by asking a colleague to deal with the odd cases, your saving and the workload for everyone are different accounts.

For a repair, make a short note of the problem, the time spent on it, the time it stopped you doing other work, and how it ended. Count the time spent making that note too. When rare repairs haven’t happened while you were watching, mark their cost as unknown, not zero.

Keep different kinds of time separate

Count the time you spend working on it yourself. Also write down any waiting that stops you doing other work. Keep the time the machine runs on its own apart, and don’t add up things that happened at the same time twice.

Say a process runs while you get on with another task. The whole time it runs isn’t automatically time you’ve lost. If it needs your attention the whole way through, that’s different from leaving it to run on its own.

This is the difference the vacuum study makes clear. The machine running for longer and the household cleaning less by hand could both be true. Getting ready and keeping an eye on it still needed their own place in the account of human work.

Read the total without hiding unknown costs

Add up four things: the setup time you’ve given to your period, the old work you still do, any new work and any waiting that stops other work. Count each part only once. Compare the total with the human time needed before automation, for the same result and period.

Here’s a made-up example in minutes, over one chosen period:

Part of the comparisonIllustrative human minutes
Old total, including its existing checks120
Setup allocated to this period20
Old work still required after automation30
New preparation, checking and upkeep, counted once40
Waiting that prevents other work10

The new total is 20 + 30 + 40 + 10 = 100 minutes. The change is 100 − 120 = −20 minutes: 20 minutes less human time in this account.

Now say the cost of rare repairs is still unknown. Then that 20-minute difference is only provisional, because a cost in the account is still open. Keep the known total and the gap together. Use a range where you have a sensible basis for one.

If the full new total is lower, the account shows a saving for that result and period. If it’s higher, it shows extra human time. If the totals match, neither way saves time in that comparison. If important costs are still unknown, write down what you’d need to find out before you claim a full saving.

Keep the benefits and optional interactions visible

Write down what the remaining work gets you: better output, convenience, easier access or less boring work. An arrangement that takes the same time can still be worth having. Describe that benefit as it is, instead of turning it into made-up minutes saved.

Look separately at optional app use or alerts. The vacuum users who switched to the button show that using a device doesn’t always mean using its app.

Before you remove useful alerts or updates, make sure you still have a way to tell whether the result you want still arrives. Less fiddling isn’t helpful if it hides failures that need fixing later. Count that fixing work if it happens.

Swap estimates for what you see as you use the automation. Note the real getting ready, checks, odd cases and upkeep, and change the account when what you see changes it. You can then explain both the work involved and the benefits you value.

What work sits around your automation?

How about your daily life?

Where do you get an automatic result but still do work around it? You might get inputs ready for a regular report, check a draft made with AI help or look after a device at home.

Which work around it is hardest to see: setting up, checking, the odd repair or someone else’s help? Write down situations where a clearer account would be useful. One good idea is enough.

A routine automation in my daily lifeWhat I want to understand
________________________________________
________________________________________

My chosen situation: ____________________

Choose one routine automation, at home or at work, to take into the exercise. You can change your choice. Choosing a situation doesn’t mean it saves time or needs replacing.

Make a complete time account

Start with your chosen situation. If you came straight here, choose one routine automation, at home or at work, whose workload you want to understand. This is a written account of the work and the result.

  1. State the result. Describe how much output there is, what counts as good enough and who is involved. If automation makes more, write down that workload apart from the work needed for the same output as before.
    Result and people: ____________________
  2. Record the old human work. Include getting ready, doing the task, checking, fixing and coordinating. Include checks you already did before automation.
    Old work and total time: ____________________
  3. Account for one-off setup. List the time spent getting the system ready, learning to use it and testing it. Choose a period, write down how many times you expect to use it, and say how much setup time you count in that period. If you spread setup across uses, say so, and don’t count its full cost again and again.
    Period, uses and allocated setup: ____________________
  4. List ongoing work. Include getting ready, coordinating, keeping an eye on things, checks, fixes, problems, upkeep, updates, breakdowns and moves to a replacement system where they apply. Write down who does each one, how often, the minutes of work each time and anything you’re unsure of. Keep routine upkeep apart from unusual repairs, without counting one event twice.
    Activities, who does them, how often, minutes and unknowns: ____________________
  5. Separate the clocks. Write down the time you work on it yourself, and any waiting that stops you doing other work. Keep the time the machine runs on its own apart, and don’t count things that happen at the same time twice.
    Working time, waiting that blocks other work and separate machine time: ____________________
  6. Include what remains or moves. Write down the steps you still do by hand, and work handed to another person or a company. Keep your own time apart from the total for everyone involved.
    Remaining work and other people’s work: ____________________
  7. Compare the totals. Add the setup time given to your period, the old work you still do, new work counted once and waiting that stops other work. Take away the old total for the same result and period. A negative difference means less human time; a positive one means more. Zero means the totals match. Use a range when you have grounds for one, and leave the cost of rare events unknown when you don’t have the information.
    New total, old total, difference and unresolved costs: ____________________
  8. Record benefits beside time. Describe better output, convenience, easier access or less boring work. Describe those benefits without making up a time saving.
    Benefits worth retaining: ____________________
  9. Replace estimates with experience. Keep short notes on getting ready, checks, problems, upkeep and moves to replacement systems. Count the time spent making the notes too. For a failure, note the problem, the minutes of work, the time it stopped other work and how it ended. If you change optional alerts or app use, keep a way to check that the result you want still arrives.
    What to observe next and how to confirm the result: ____________________

Ask Jing

Become a Jingius.

Count the work around one automation

Loading interactive tool…

Your next step

Finish with a time comparison for a stated result and period, or an account that isn’t complete yet, with the missing costs clearly named. Keep the benefits alongside it. This gives you a clear picture of what the automation asks of you and what you get from it.

A note on the evidence

The sources look at different settings and different results. The vacuum study recorded less cleaning by hand, but left the work around it out of its chart of cleaning time. The smart-home report describes people’s experiences, while the writing experiment measures how people did on tasks they were given.

These findings support looking at both the work taken away and the work that’s left. They don’t give you a general rate for upkeep, or a full lifetime saving for your own automation.

How we prepared this module

This module links a field study in homes, a paper on how automation works, research on how people live with smart homes and a writing experiment. The made-up time comparison shows how the sums work. The written exercise helps you organise your own notes, estimates and open questions.

Sources