The Routine-Work Tax: Why Your Best People Spend Half Their Day Below Their Pay Grade

You hired a senior analyst for judgement. You pay a specialist to solve problems only they can solve. You promoted a team lead because their thinking moves the business. Then you look at how they actually spend the week, and half of it is data entry, chasing a colleague for a status update, copying figures from one system into another, and answering the same question they answered last Tuesday.

That gap has a price, and almost nobody puts it in a budget. When a skilled, well-paid person does routine, low-value work, the company pays an expert salary and receives clerical output for those hours. Multiply that across a department and a year, and it becomes one of the largest unmanaged costs on the payroll. It is a tax you pay every month without ever seeing the invoice.

This piece names that tax, models what it costs in euros, explains why the copilots and point tools most companies reach for do not remove it, and gives a practical way to pay it down. The argument is simple: the fix is not another tool that makes routine work faster, and it is not more headcount. It is leverage, so the routine work stops landing on your best people in the first place.

What the Routine-Work Tax Is

The routine-work tax is the difference between what you pay for a skilled hour and what you actually get back when that hour is spent on routine work. It is not fraud, waste, or laziness. It is a structural mismatch: expensive capacity pointed at cheap tasks, repeated so often that everyone stops noticing.

A precise definition

• Expert input, routine output - You pay for judgement, expertise, and problem-solving, and for a large share of the week you receive keystrokes, lookups, and copy-paste instead.
• Invisible by design - There is no cost centre called “senior people doing junior work.” The cost is buried inside salaries you have already committed to, so it never triggers a review.
• Below the pay grade - The work itself is often necessary, but it does not require the person doing it. A form still needs filling; it does not need a specialist to fill it.
• Regressive in the worst way - The tax rate rises with salary. The more you pay someone, the more expensive each routine hour becomes, so your most valuable people carry the heaviest load.
• Compounding with growth - Every time you hire or promote, you attach the same routine load to a bigger salary, so the absolute cost grows year after year.

What counts as routine work

Routine work is not defined by the department. It is defined by whether the task needs the specific person, or just a pair of hands that knows the context. The usual suspects appear in every function.

None of these tasks is glamorous, and none of them is optional. The question is not whether they get done. It is whether your most expensive people should be the ones doing them.

Where the Hours Actually Go

If the tax were small, it would not be worth naming. The evidence says it is not small. Independent studies using different methods keep landing in the same range: routine and coordination work consumes roughly half the knowledge-worker day, and genuinely skilled work gets what is left.

The independent evidence

• Half the day on low-value tasks - A 2025 survey of 2,000 knowledge workers, commissioned by HP and run by Talker Research, found people spend 51 percent of their working hours on tedious, low-value tasks¹.
• Only a quarter on skilled work - Asana’s Anatomy of Work Index finds the average person spends about a quarter of their time on skills-based work and roughly 60 percent on “work about work”: chasing information, switching between apps, and managing shifting priorities².
• Email and search dominate - The landmark McKinsey Global Institute analysis put email at 28 percent of the working week and searching for internal information at a further 19 percent, so nearly half the week went to two coordination activities alone³.
• The task breakdown is telling - In the HP survey, writing emails took 31 percent, data management 25 percent, catching up on team communications 22 percent, searching for and organising files 18 percent, and managing calendars 16 percent¹.
• Duplicative work is enormous - Asana estimates the average knowledge worker loses 209 hours a year to duplicated work and 352 hours a year to talking about work rather than doing it².
• Leaders see it too - 76 percent of IT decision-makers in the HP survey agreed that employees waste too much time on menial work¹.

The burnout multiplier

The routine-work tax is not only a cost line. It is also a retention risk, because the people best equipped to do valuable work are the ones most frustrated by not getting to.

• Repetition drives burnout - 85 percent of workers in the HP survey said repetitive tasks contribute to burnout¹.
• Technology adds friction - A third of workers said their current work technology contributes to feelings of stress, and one in three had considered quitting over frustrating tools¹.
• The work-about-work tax hits morale - 63 percent of Asana respondents reported experiencing burnout in the past year, with coordination overhead a leading cause².
• Your best people notice first - High performers have the clearest sense of what their time is worth, so they feel the mismatch most sharply and leave for it soonest.

The pattern is stable across studies and years. Now it is worth turning the percentages into money, because that is the number your board will act on.

Modelling the Euro Cost

Percentages do not move budgets; euros do. The model below is deliberately conservative and every input is stated, so you can rebuild it with your own numbers. It is illustrative, not a claim about any specific company.

The inputs

• Starting salary - The average gross salary for full-time employees in Germany was roughly 59,100 euros a year in the StepStone Gehaltsreport 2026, with the Destatis median around 54,000 euros^9,10. Skilled specialists sit above the average.
• Fully-loaded cost - Add employer social contributions, roughly 20 to 25 percent in Germany, plus workplace and overhead. A 60,000 euro gross salary lands near 70,000 to 75,000 euros fully loaded. We use 70,000 euros as a round, conservative figure.
• Routine share - The studies suggest 50 to 60 percent of time on low-value work^1,2. We deliberately use 40 percent, well below the evidence, to keep the model unarguable.
• The tax per person - 40 percent of 70,000 euros is 28,000 euros a year of expert salary buying routine output, for a single skilled employee.

How it scales

The tax is linear in headcount and in salary, which is why it grows quietly as the company succeeds. The table models the annual tax at a 40 percent routine share and a 70,000 euro fully-loaded cost.

Sensitivity: it gets worse before it gets better

The 40 percent figure is intentionally low. If you use the routine share the studies actually report, the tax roughly tracks toward the evidence. Here is the per-person tax at a 70,000 euro fully-loaded cost across different routine shares.

The number is real, it is large, and it compounds. The natural next question is why the tools most companies have already bought have not made a dent in it.

Why Copilots and Point Tools Do Not Remove It

Most companies have already spent money on this problem. They rolled out a general copilot, bought a point tool for one team, and wrote some automations. The routine-work tax barely moved. The reason is structural, not a failure of any single product.

Speeding up a task is not owning an outcome

• Copilots accelerate keystrokes - A general assistant drafts a faster email or summarises a document, but the person still opens the app, checks the result, and does the next step. The workflow still belongs to the human.
• Point tools solve one slice - A best-in-class tool for one team optimises that team’s task and stops at its own boundary. The handoffs between tools, where routine work actually lives, stay manual.
• RPA is brittle - Scripted automation follows fixed steps and breaks the moment a screen, field, or exception changes, so a skilled person is pulled back in to fix it.
• The obligation does not move - All of these make the task cheaper per unit. None of them removes the human’s ownership of the end-to-end result, which is where the tax lives.

The context problem

Routine work only looks routine from the outside. From the inside it is full of the small judgements that make automation guess wrong and hand everything back to a person. This is why generic tools stall.

• Generic AI knows the internet, not your company - It has never seen your approval rules, your exception list, or how you define a qualified account, so it cannot act the way your staff would.
• The knowledge lives in people - The reasoning behind decisions usually sits in a few experienced heads, not in any document, so a tool with no memory of it stays shallow.
• Handoffs need a shared brain - Moving an outcome across CRM, ERP, email, and a spreadsheet requires one system that understands all of them and the rules connecting them.
• Hype outruns capability - Gartner estimates only around 130 of the thousands of vendors marketing agentic AI are genuinely agentic, a practice it calls “agent washing”⁵.

If speeding up tasks does not remove the tax, the alternative has to change what lands on the person in the first place. That is a question of leverage.

Leverage, Not Headcount

There are only three ways to deal with routine work: do it with expensive people (the tax you pay now), hire cheaper people to absorb it (more headcount, more coordination, the same work), or take it off people entirely (leverage). The third is the only one that scales, and it is now practical.

What leverage actually means

• Each person keeps working as usual - Nobody changes tools or learns a new platform. The difference is what stops arriving in their inbox and their queue.
• The routine work no longer hangs on them - Defined, repeatable outcomes get owned by an AI employee, so the human is freed for the parts that need a human.
• Grounded in a Company Brain - The AI employee works from a persistent store of your rules, definitions, and exceptions, so it acts the way an experienced colleague would rather than guessing.
• Connected to the real systems - It operates across the tools you already run: email, Teams, SharePoint, CRM, and ERP, moving outcomes across them instead of stopping at one boundary.
• Human in the loop for exceptions - It handles the routine 80 percent and escalates the genuinely uncertain cases, so oversight stays where it adds value.

Why the timing is right

• The models finally understand context - McKinsey attributes the jump in automation potential to generative AI’s ability to understand natural language, which underlies work activities worth 25 percent of total work time⁴.
• Automation potential is now most of the day - McKinsey estimates current AI could automate activities that absorb 60 to 70 percent of employees’ time, up from earlier estimates of half⁴.
• Agents are entering real software - Gartner projects 40 percent of enterprise applications will feature task-specific AI agents by 2026, up from less than 5 percent in 2025⁶.
• Routine decisions are next - Gartner expects at least 80 percent of governments to deploy AI agents to automate routine decision-making by 2028, a signal of how mainstream the pattern is becoming⁷.
• The skills base is shifting under everyone - The World Economic Forum expects 39 percent of core skills to change by 2030, so capacity for higher-value work is the binding constraint⁸.

Leverage is the goal. The rest of this piece is about how to reach it deliberately, one workflow at a time, without betting the company on a moonshot.

The Pay-Down Playbook

Paying down the routine-work tax is not a big-bang transformation. It is a sequence of focused moves, each of which removes one routine load from one team and proves the value before the next. Here is the practical version.

Step by step

1. Find the most expensive routine line - List the workflows where your highest-paid people spend the most time on the lowest-value tasks. Rank by salary times hours, not by how annoying the task feels.
2. Baseline the tax in euros - Measure current hours, the fully-loaded cost of the people doing them, and the error and delay costs. This is the number you will pay down and report against.
3. Capture the context - Sit with the people who do the work and record the rules, exceptions, and reasoning into a Company Brain. This is where routine work becomes automatable rather than merely repetitive.
4. Connect the real systems - Give the AI employee access to the tools the workflow already runs on: email, Teams, SharePoint, CRM, ERP. No new platform for the team to learn.
5. Run in parallel first - Let the AI employee handle the routine outcome alongside the human for a few weeks. Compare results, tune the rules, and build trust before you hand over.
6. Define the human-in-the-loop line - Decide which cases the AI handles autonomously and which it escalates. Keep oversight on exceptions and anything high-risk.
7. Measure reclaimed hours - Track hours returned to skilled people and where those hours now go. Reclaimed time only counts if it lands on higher-value work.
8. Expand to the next line - Once one workflow is stable, reuse the same Company Brain and connections for the next. The second deployment is faster than the first.

What good and bad look like

How Superkind Fits

Superkind builds AI employees for companies. The positioning is deliberately narrow: not a copilot that makes tasks faster, but an AI employee that takes a routine outcome off your people entirely, grounded in your company’s context and connected to the systems you already run.

What the AI employees do

• Owns routine outcomes - The AI employee runs a defined workflow end to end, so the result happens without a skilled person babysitting each step.
• Grounded in a Company Brain - It works from a persistent store of your rules, definitions, and exceptions, so it applies your logic rather than generic internet knowledge.
• Connected to your real stack - It acts across email, Teams, SharePoint, Salesforce, HubSpot, SAP, and other CRM and ERP systems as one layer over what you already use.
• Handles exceptions with oversight - It runs the routine cases autonomously and escalates the uncertain ones, keeping humans on the decisions that need them.
• People keep working as usual - Nobody adopts a new platform; the routine work simply stops landing on them.
• Live in weeks, not months - First workflows go into production quickly, then get sharper as the team gives feedback.
• Learns from your team - Corrections and feedback flow back into the Company Brain, so the AI employee improves the way a new hire would.
• Built around your process - The starting point is your existing workflow, not a template you have to bend your company to fit.

The article holds up without the product: name the tax, model it, and pay it down with whatever removes routine work end to end. Superkind is simply built to do exactly that.

Decision Framework: How Much Are You Paying?

Not every company should act on this today, but most that employ skilled people should at least measure it. Use these signals to decide where you stand.

Frequently Asked Questions

Related Articles

• The AI Productivity Paradox: Why Individual Wins Do Not Add Up to Business Value
• Hiring Freeze, Not Layoffs: How to Grow Output While Headcount Stays Flat
• The Last-Mile Problem: Why AI Pilots Impress in the Demo and Die at Handoff
• The End of Labor Arbitrage: How AI Employees Outwork the BPO Model
• Time-to-Productivity: How a Company Brain Makes New Hires Productive in Days, Not Months

Henri Jung

Co-founder of Superkind, where he helps SMEs and enterprises deploy custom AI agents that actually fit how their teams work. Henri is passionate about closing the gap between what AI can do and the value it creates in real companies. He believes the Mittelstand has everything it needs to lead in AI - it just needs the right approach.