The Hidden Tax of "I'll Just Do It Myself"

It is 7:40pm. A VP of operations has one thing left before she can close her laptop: prep for a 9am with a key account. She opens Slack to find the last thread. Then HubSpot for the renewal date. Then a shared drive for a deck someone half-finished. Then her inbox again, because one number does not match. Forty minutes later she has a clean one-pager and a cold dinner. She could have asked someone on her team to pull it all together. But explaining what she needed would have taken longer than just doing it.

That last sentence is the tax. Every week, senior people quietly absorb hours of small operational work because handing it off feels slower than finishing it. The work is real, it adds up, and almost nobody counts it.

The pattern has a name, and you are probably running it

“I’ll just do it myself” feels responsible. The jobs are small. The context already lives in your head. Delegating means a round of questions, a half-right draft, and a correction pass before anything is usable. Doing it yourself skips all that. So you do it. Then you do it again tomorrow, and the day after.

Each decision makes sense on its own. The trouble is that you make it a hundred times a month, and the total is a part-time job you never agreed to take. That is arithmetic, and better time management will not fix it.

What the tax actually covers

Most of it hides inside four jobs that rarely show up on anyone’s goals:

• Chasing context. Finding the latest version of a file, the real status of a project, the decision made in a meeting you missed, or the one person who knows why a number changed.
• Writing follow-ups. Turning a call into a recap, a list of next steps, and three nudges to the people who owe you something.
• Pulling information from systems. Logging into HubSpot, Jira, a drive, and your inbox to assemble one view that no single app gives you.
• Preparing meetings. Stitching together an agenda, the latest figures, and a short brief so the room is not starting cold.

None of these move a strategy forward on their own. Asana calls this category work about work: the coordination, status chasing, and duplicated effort that sit around the real job. Its own research puts the scale at 60% of a knowledge worker’s time. It is easy to do, easy to justify, and nearly invisible until you add it up.

A team lead feels this most on Friday afternoons, when the weekly status update means opening six tabs to write four sentences nobody reads closely. An ops manager feels it every time a “quick question” in Slack turns into a twenty-minute dig through last quarter’s tickets.

How big the bill gets

The numbers are worse than they feel, partly because each task is small enough to ignore.

Start with search. APQC found that knowledge workers lose about 8.2 hours a week looking for, recreating, or duplicating information. That is a full workday, gone before anyone writes a strategy doc or closes a deal.

Then add the cost of jumping between apps to do the searching. Researchers writing in Harvard Business Review tracked workers switching between apps about 1,200 times a day. The reorienting after each switch cost them just under four hours a week. Every toggle is cheap. The habit is not.

Meeting prep has its own tell. Microsoft’s Work Trend Index found frequent interruptions through the workday and a 122% spike in PowerPoint edits in the final ten minutes before a meeting. That last-minute scramble is what happens when the person with the context is also the only person who can assemble it, and they run out of time to do it earlier.

The load is heaviest on the people you would least expect to be doing admin. Gallup found that 97% of US managers still carry individual contributor responsibilities, spending a median of 40% of their time on that work. Deloitte put a similar share, close to 40% of managers’ time, into firefighting and administrative work. So the senior person who was promoted to set direction spends two days a week on operational jobs they were supposed to hand off.

It also lands unevenly across a team. Economist Linda Babcock and her co-authors studied what they call non-promotable tasks for their book “The No Club,” and found that women spend roughly 200 more hours a year than men on this kind of low-recognition work. Whoever is reliable and good at the plumbing tends to get more of it. That quietly shapes who has time for the work that gets noticed.

Why it keeps happening: a trust and handoff problem

Senior people hold these jobs because delegating them often costs more than doing them.

To hand off meeting prep, you have to explain which account, where the files live, what “the latest numbers” means, which thread has the real decision, and what good looks like. That explanation is context switching, admin, and coordination overhead, now for two people instead of one. Then you wait. Then you review a draft that is 70% right. Then you fix the last 30%, which was the hard part all along.

For a task you do in fifteen minutes, that handoff can take forty. So you keep it. You may trust your team, but moving context is slow and lossy, and the risk of a wrong output landing in front of a customer sits with you. HBR has argued that collaboration overload is usually a symptom of a deeper organizational problem rather than a personal failing. The same is true here. When small jobs pile onto senior people, the real issue is that your organization has no cheap way to move context out of a person’s head and to someone, or something, that can act on it.

Once you see it as a handoff problem, the fix gets clearer: make moving context nearly free.

What changes when an agent is in the loop

This is where we have spent our time at Autohive.

The expensive part of delegation is moving context. An Autohive agent can read from your files and live systems such as HubSpot, Jira, Slack, Microsoft 365, Google Workspace, and GitHub. Finding information becomes a plain-language request, answered against your real sources rather than a stale copy. We wrote more about chasing context across fragmented systems.

Meetings get captured instead of re-created. Autohive Meetings records the transcript, decisions, and action items, then lets you pass that context to a specialist agent. The recap and follow-ups that used to eat your evening become a first draft waiting for your edit.

Recurring work stops needing a human trigger. Scheduled recurring jobs can run the weekly status check, pull the reporting, and push downstream updates. That covers the part that was always the same, so the team lead’s Friday update writes itself and waits for a glance.

Interruptions also drop because people have somewhere else to ask. In a shared workspace, your team can put questions to a grounded custom agent instead of pinging the one senior person who knows. The ops manager stops being a human search engine.

Humans keep judgment, approval, prioritization, and control over permissions. Agents handle collection, extraction, synthesis, drafting, and repeatable execution. The meeting brief gets built for you, and you decide what to say in the room. Our guide on deciding what to delegate to an AI agent goes deeper on that line. Software alone will not fix the problem until you start changing the underlying workflow. Another app to toggle into just adds to the pile.

Where to start this week

Start with one honest week of data and one small handoff. A reorganization can wait.

• Track the small jobs for five working days. Every time you chase context, write a follow-up, pull data from a system, or prep a meeting, note what it was and roughly how long it took.
• Find the one you repeat most. Choose the most frequent task. Repetition is where the tax compounds.
• Write down what you actually do. The steps, the sources, and what a good result looks like. This is the context that usually never leaves your head.
• Hand off a narrow first version. Give an agent that one task with those exact sources and that definition of done. Keep it small on purpose.
• Keep review and judgment. Check the output, approve what ships, and tighten the instructions where it missed. You are supervising a draft, and the decision stays yours.

The work you reclaim is the forty minutes before a 9am and the hour that disappears into a status update. Over a year, that decides whether a senior person sets direction or spends two days a week as a very expensive search engine. Start by counting it, because you cannot hand off what you have never named.

FAQ

What is work about work?

Work about work is the coordination and admin around the actual job: chasing status, searching for files, re-creating information, duplicating effort, and sitting in meetings to align. It feels productive and often is not. Asana’s research uses the term. APQC’s finding that knowledge workers lose about 8.2 hours a week looking for, recreating, or duplicating information gives a sense of the scale.

How many hours can this cost?

More than most people guess, because each task is small. APQC puts searching for, recreating, and duplicating information at roughly 8.2 hours a week. Harvard Business Review found workers lose close to four hours a week to switching between apps. Gallup and Deloitte both found managers spending around 40% of their time on individual contributor and administrative work. The jobs feel like minutes and add up to days.

Can an AI agent handle judgment calls?

An agent is good at collecting context, pulling data from your systems, synthesizing it, and drafting repeatable output. Judgment, priorities, approval, and the decision about what a customer sees stay with you. Let the agent do the assembly and keep the call for yourself.

Where should I start?

Track the small operational jobs for one week, pick the one you repeat most often, and hand off a narrow first version of it to an agent with your real sources and a clear definition of done. Keep review and final judgment. One task done well beats a grand automation plan you never ship.