The personal workflow analysis: what watching a real workday reveals about automation

Ask someone how they spend their day and they will describe the work that matters: client meetings, decisions, problem-solving, strategy.

Now record that same day, the actual sequence of screens, clicks, and keystrokes, and a different picture emerges. Between the meaningful work sits hours of something else entirely.

The gap between what people report and what they actually do is not dishonesty. It is human nature, and it has been measured. Time-diary research published in the US Bureau of Labor Statistics’ Monthly Labor Review compared people’s estimates of their working week against diaries of what they actually did, first in 1994 and again in 2011. Estimates ran 5 to 10 per cent high on average, and the bigger the claimed week, the bigger the gap: in the 1994 data, people who said they worked 55 to 59 hours a week were about 10 hours over what their own diaries showed. That is government-published time-use research, not consulting folklore. Memory is simply a poor instrument for measuring time.

We remember the work that engages us. We forget the work that does not. And it is the forgettable work, the copying, reformatting, re-entering, and repeating, where the biggest automation opportunities live.

You cannot find these patterns by asking. You can only find them by watching.

Why do interviews miss the work that matters?

Interviews miss it because people describe their role, not their routine. Every consulting engagement in history has started with some version of “walk me through your typical day”, and the answers are always about the job, never about the keystrokes.

A conveyancer will tell you about managing settlements, liaising with banks, and advising clients. They will not mention the forty minutes spent copying the same property details from an email into their practice management system, then into the contract, then into the settlement statement. That is not how they think about their day. It is just what happens between the parts they think about.

A trades business owner will tell you about quoting jobs, managing crews, and keeping clients happy. They will not mention that every quote involves re-entering the same client details into three different places, or that their follow-up process is a mental note that gets lost by Thursday.

These are not small inefficiencies. They are hours per week, but they are invisible to the person doing them because they feel like “just doing the work”.

A recorded workday makes them visible. Not through interpretation or estimation, but through direct observation of what actually happens, in what order, for how long.

What does observation research actually show?

That a working day is far more fragmented than anyone living it would report. The strongest evidence is a peer-reviewed observational study: in 2005, Gloria Mark and colleagues at the University of California, Irvine shadowed 24 information workers for roughly 26 hours each, timing every activity to the second. On average, people spent 11 minutes and 4 seconds on one sphere of work before switching or being interrupted. Fifty-seven per cent of those working spheres were interrupted. And once interrupted, getting back took an average of 25 minutes and 26 seconds, with 2.26 other tasks handled in between.

Nobody describes their own day in those terms. A stopwatch did. The study is two decades old, which is part of why it is useful: the fragmentation it measured is structural to office work, not a smartphone-era affliction. And the method is the point. The researchers did not ask. They watched.

What becomes obvious once you watch?

Three patterns, in almost every business. The specific findings vary, but these appear every time.

The same information moves between systems repeatedly. An enquiry arrives by email. The details go into a CRM. The CRM informs a proposal. The proposal becomes a project brief. Each step carries most of the same content in a different format. In healthcare, a referral becomes an intake note becomes a treatment plan becomes a progress note, each document inheriting from the previous one and adding something new. The time spent re-entering and reformatting that inherited information is pure overhead.

Predictable events trigger identical sequences. A new enquiry arrives and triggers the same five steps. A quote is accepted and triggers the same eight. A project completes and triggers the same follow-up. You do not experience these as repetitive because each instance involves a different client. But the structure is identical every time, and structure is what automation runs on.

People act as connectors between systems. The invoice generated from data that already exists in the project system. The report that pulls numbers from a spreadsheet into a template. The calendar update that reflects information already captured elsewhere. No creativity, no expertise, no decision-making. Just a human bridging a gap between two systems that could talk to each other directly.

Once you see these patterns in a recording, you cannot unsee them. And each one maps directly to a specific automation opportunity with a measurable time saving.

Where does this fit in workflow mapping?

The personal workflow audit is the ground-level instrument of workflow mapping, the first step of how we work. Our model runs in three steps: workflow mapping to learn how your business actually runs, AI design to identify what is worth automating, then build. We build it, or coach your team to. Either way, you own the result.

Mapping works in both directions at once. A strategic AI audit looks at your business from the top down: operations, customer journey, competitive position. A recorded workday works from the bottom up. It watches one person’s actual day and finds the friction that strategy-level thinking misses. The top-down view finds the opportunity to automate your entire follow-up process. The bottom-up view finds the twenty minutes your office manager spends every morning copying yesterday’s bookings into a spreadsheet that nobody has questioned in four years.

The recording also does something estimates cannot: it hands the AI design step real numbers. “About an hour a day on email” is a guess. Fourteen minutes re-keying job details across three systems, four times a day, is a measurement, and the difference decides whether you design automation around what people believe or around what actually happens. When we run this as a paid Workflow audit, the deliverable is a written report with specific, prioritised recommendations and a clear next step. Not a slide deck, and not a proposal for more work. What a good audit actually delivers shows what that looks like on a real engagement.

What does watching miss?

Plenty, and pretending otherwise would undercut the method. A recorded day is a sample of one: the end-of-month invoice run, BAS time, and renewal season will not show up in a random Tuesday, so the biggest seasonal pain can be invisible to a single recording. People also behave differently when they know they are being observed, at least at first; the effect fades, but it never fully disappears, and we do not record covertly to get around it. Recording someone’s screen without their knowledge is surveillance, not analysis. Screen-level repetition can also hide judgement: the valuer re-typing figures may be checking them, and the check may be the actual job. And observation finds friction inside the process as it exists. It cannot tell you whether the process should exist at all.

That is why watching is one instrument, not the whole method. Interviews remain in the kit because they are good at the things recordings are bad at: intent, exceptions, and what the work means. They are just unreliable about time. The map is only honest when the two are read together.

Perth AI Consulting maps real workflows by observation, not interviews alone, then designs and builds the workflow and admin automation that follows. If you want to know where your team’s hours actually go, start with a conversation.