Automatic Tracking: What It Captures and What It Misses

Automatic capture solves the biggest problem in time tracking — that recording takes effort and therefore does not happen. It introduces a different one, and whether it is worth the trade depends entirely on how it is configured and who sees the output.

What it actually captures

Application and document activity. Which program was in focus, which file, which browser tab, for how long.

For a vendor-side perspective on mouse jiggler detection software, see this resource from Monitask.

Calendar entries. Meetings, as scheduled rather than as they happened.

Activity in connected tools. Commits, tickets moved, messages, documents edited.

Idle detection. Periods without input.

Location, on some mobile tools.

Notice the shape of this list. It measures where your attention appeared to be, inferred from device signals.

What it misses

Thinking. Reading a printout, working something out, staring at a wall. Registers as idle.

Phone calls away from the keyboard.

Conversations. The five minutes that solved the problem.

Anything away from a screen.

And intent. A document open for four hours looks like four hours of work on it. It may have been open while you did something else entirely.

The consistent bias: rewards screen presence, misses everything else. See why time data is almost always wrong.

Where the line is

This is the question that determines whether automatic tracking is fine or is a problem.

Who sees the raw capture?

If it goes to the individual as a suggestion they confirm or correct, it is a recording aid. The person remains the source of truth, and the detail stays with them.

If it goes to a manager as a record, it is monitoring — regardless of what the product is called. See the line between timekeeping and surveillance.

And activity scores, idle percentages and application breakdowns shared upward are monitoring outputs, even when the same tool also keeps a timesheet.

Using it well

Suggestion, not record. The system proposes; the person confirms. This preserves both the friction reduction and the accuracy.

Keep the raw detail with the individual. The manager sees "six hours on the Henderson project", not the application log behind it.

Let people edit before anything is submitted, without asking permission.

Do not capture what you do not need. Application-level is usually sufficient; document titles and browser URLs are frequently more than the purpose requires and carry more personal information.

Say what is captured, plainly. Ambiguity here is read as concealment, and it converts a helpful tool into a suspected one.

And set retention. Detailed activity capture accumulates a great deal of personal data quickly. See time records as employment records.

Where it genuinely helps

Catching what you forgot. The half hour you did not log.

Prompting at the right moment. "You have been in this document for 45 minutes — log it?"

Reconstructing a day you failed to record, which is better than reconstructing from memory alone.

Calendar import for meetings, which is low-risk and saves real effort.

And for solo workers, removing the single biggest reason tracking lapses — that starting a timer is one more thing to remember.

Where it causes problems

Presented as accurate. It is a signal, not a measurement, and a report built on it inherits the screen-presence bias without saying so.

Shared upward in detail. See above.

Used to question gaps. "You were idle for forty minutes" is a monitoring conversation, and one of them ends honest recording for good.

Applied to work that is not screen-based, where it measures almost nothing relevant.

And treated as a substitute for the purpose statement. The tool does not decide what the data is for; you do. See what time tracking is for.

For individuals

The trade is different when nobody else sees it. For broader independent background, see ILO guidance on working time.

Automatic capture is genuinely useful solo — it removes the friction that ends most personal tracking.

And the bias still applies. Your logged hours will over-represent screen time and under-represent thinking, which matters if you use the data for quoting.

Correct it rather than trusting it. Treat the capture as a draft of your day.

And do not score yourself on activity percentages. Same trap as monitoring, aimed inward. See time tracking for freelancers.

Questions for a vendor

What exactly is captured, at what granularity?

Who can see the raw capture — only the individual, or also managers?

Can the individual edit before anything is shared?

Can specific capture types be disabled, and is that verifiable?

What is the retention, and can we set it?

And can it run in suggestion mode rather than recording directly? If not, it is a monitoring product.

The short version

It captures screen presence and infers work from it — thinking, calls and conversations register as nothing.

Who sees the raw capture is the whole question. To the individual as a suggestion: recording aid. To a manager as a record: monitoring.

Suggestion mode preserves both the friction reduction and the accuracy.

Capture less than the tool offers, because detail you do not need is personal data you are holding.

And it is a signal, not a measurement — reports built on it inherit the bias whether or not they say so.