Here’s a number: $174,720.
That’s how much a 40-person professional services firm is likely spending every year on manual work that doesn’t require a human brain. Not in some theoretical model. In actual loaded labor costs, calculated from actual survey data about how people actually spend their days.
Nobody tracks this number. It doesn’t appear on any line item. There’s no vendor invoice for it. It just bleeds out, week after week, in 15-minute increments across your entire team — status update emails, manual data entry, chasing approvals through Slack threads, reformatting the same document into three different templates because your systems don’t talk to each other.
I want to walk through the math, because nobody ever does. And then I want to ask a question that might change how you think about your next hire.
The Math Nobody Does
Let me show my work. This isn’t complicated, but apparently it’s rare.
According to the Asana Work Index, employees spend roughly 60% of their time on “work about work” — status updates, searching for information, chasing approvals, switching between tools, and other coordination overhead that doesn’t directly produce value. For a knowledge worker putting in 40 hours a week, that’s 24 hours spent on work that isn’t the actual work.
Not all of that is automatable. Some coordination is genuinely necessary. But the portion that’s pure manual process — data entry, document reformatting, status reporting, approval routing — consistently lands around 7 hours per person per week in professional services firms, based on workflow audit patterns I’ve seen and industry benchmarks from McKinsey’s research on knowledge worker productivity.
Seven hours per person per week. That’s the number that matters for the math I’m about to do, and it’s the kind of copy-paste work that makes your best people quietly update their LinkedIn profiles.
Now take a 40-person firm. Not everyone’s doing this work — your C-suite isn’t reformatting spreadsheets (hopefully). But conservatively, half your team is. That’s 20 people.
20 people x 7 hours per week x 52 weeks = 7,280 hours per year.
At a blended loaded cost of $40 per hour — and that’s conservative for professional services, where loaded costs often run $50-$65 — that’s $291,200 annually in labor dedicated to tasks that require a pulse but not a brain.
Now, can you eliminate all of it? No. Some of it requires human judgment that’s genuinely human. Some of it involves edge cases that would cost more to automate than to just do manually. But based on workflow audit patterns, roughly 60% of that manual work is recoverable through a combination of automation, AI workflow design, and basic process cleanup.
Sixty percent of $291,200 is $174,720.
That’s the number. That’s what’s sitting on the table.
Where the Hours Actually Go
I ran a professional services startup. I know exactly where this time goes because I watched it happen for years before I understood what I was looking at.
Status update emails. Someone finishes a task. They email their manager. The manager emails the project lead. The project lead updates the client spreadsheet. The client asks for a weekly summary, which someone manually compiles from all those email threads on Friday afternoon. Three to four people touched information that could’ve flowed through a system automatically. Multiply by every active project.
Data entry between systems. Your CRM doesn’t talk to your project management tool. Your project management tool doesn’t talk to your invoicing system. So someone — usually your most operationally capable person, the one you can least afford to waste — spends hours every week being a human API. Copying client details from one system to another. Reconciling numbers between spreadsheets that should’ve been the same spreadsheet.
Chasing approvals. A proposal needs sign-off from two partners. It sits in someone’s inbox for three days. A follow-up email goes out. A Slack message. A hallway conversation. The approval takes 30 seconds. The chasing took 45 minutes across four people’s calendars. Scale that across every approval in the business — expenses, deliverables, contracts, time-off requests.
Reformatting documents. The client wants the report in their template. The internal team works in a different template. The finance team needs the data in a third format. So someone manually reformats the same information into three different containers. Every single time. This is the kind of work that feels small on any given Tuesday but consumes hundreds of hours per year.
None of this is glamorous. That’s the point. Nobody builds a dashboard to track “hours spent reformatting deliverables.” Nobody puts “reduced approval-chasing time” on a KPI slide. The waste is real, it’s constant, and it’s invisible precisely because it’s distributed across the entire team in small increments that nobody aggregates.
Why Nobody Tracks This
I spent years running a business without ever calculating this number, and I’m someone who thinks in P&L statements. So I understand why it doesn’t get tracked. There are structural reasons, not just laziness.
There’s no line item. Your accounting system has a line for salaries, benefits, software, rent. It does not have a line for “hours your team spent on tasks that could be automated.” The cost hides inside the salary line, indistinguishable from productive work.
It’s distributed. No single person is spending 40 hours a week on manual processes. It’s 45 minutes here, an hour there, across 20 people. Nobody experiences it as a $174K problem because nobody experiences more than their own slice.
It’s normalized. “That’s just how we do things.” The most expensive phrase in operations. When a process has existed for three years, it stops being questioned. It becomes infrastructure. The team works around it. New hires learn the workaround during onboarding and assume it’s intentional.
The people who see it can’t fix it. Your ops manager knows exactly how much time gets wasted on manual reporting. They’ve probably mentioned it. But fixing it requires buy-in, budget, and time — three things that are already stretched thin. So the workaround persists, and the waste compounds.
The Hiring vs. Fixing Decision
Here’s where the math gets interesting, and where most companies make the wrong call.
When a team is overwhelmed, the default response is hiring. And hiring feels decisive. You post the job, you interview, you extend an offer, you announce the new team member at the all-hands. It feels like progress.
But think about what you’re actually buying. A new hire at a 40-person firm — let’s say mid-level, $70K salary, $90K fully loaded with benefits and overhead — gives you roughly 1,800 productive hours per year after PTO, onboarding ramp, and the inevitable productivity dip of the first six months.
Two hires gets you 3,600 hours and costs roughly $180K-$200K per year, every year, with annual raises baked in.
Meanwhile, a process audit and workflow automation project — the kind that actually addresses the 7,280 hours of manual work — typically runs $30K-$60K in implementation costs, plus $8K-$15K per year in maintenance. One-time build, ongoing but manageable cost.
The automation doesn’t replace 7,280 hours. It recovers 60% of them — roughly 4,370 hours — and it does it without adding headcount, without onboarding time, without PTO, and without the salary increasing 3-5% every year.
I’m not saying never hire. I’m saying check whether you need more people or fewer broken processes. Because those are very different problems with very different solutions, and one of them costs a fifth of the other.
What a Process Audit Actually Looks Like
I get specific here because “process audit” sounds like consultant-speak, and I want you to know what this actually involves. I’m drawing from my experience running a small business, where I eventually learned — too late and too expensively — that most of our operational pain came from processes nobody bothered to examine.
Step one: time tracking, but honest. Not the time tracking you do for billing. The time tracking where you ask people to log what they actually do for two weeks, including the stuff they’re embarrassed about. “Spent 40 minutes reformatting the Henderson report because our template doesn’t match their template.” That kind of thing. This is uncomfortable and essential.
Step two: process mapping. For every workflow that shows up as a time sink, draw the actual flow. Not the flow in the employee handbook — the actual flow. Who touches it? Where does it stall? What’s the handoff mechanism? Where does information get re-entered? You’ll find that most multi-step processes have 2-3 steps that exist only because two systems don’t talk to each other.
Step three: categorize. For each time sink, ask: Is this genuinely complex and judgment-dependent? Is it repetitive and rule-based? Is it somewhere in between? The rule-based work is your automation target. The complex stuff stays human. The in-between stuff usually needs process cleanup before it needs technology.
Step four: do the dollar math. For each automation target, calculate: hours per week x people involved x loaded hourly rate x 52 weeks. Then estimate what percentage is realistically automatable. Be conservative. Use 50-60%, not 90%. Anyone promising you 90% recovery is selling something.
This is not a six-month engagement. For a 40-person firm, a focused process audit takes 2-4 weeks if you know what you’re looking for.
The 60% Recovery: What’s Realistic and What’s Not
I want to be transparent about the 60% number because it matters.
Sixty percent is not a guarantee. It’s an estimate based on workflow audit patterns in professional services firms. Some processes yield 80-90% automation — high-volume, rule-based data entry is the obvious example. Some yield 20-30% — anything involving nuanced client communication or subjective judgment calls.
The 60% blended number assumes you’re working across a mix of process types and that you’re realistic about where AI and automation actually help versus where they create new problems.
Here’s what tends to be highly recoverable: data transfer between systems, standard report generation, first-draft document creation from templates, approval routing, status aggregation, and scheduling. These are the 80-90% recovery tasks.
Here’s what tends to be less recoverable: client relationship management (the actual relationship part, not the CRM data entry), creative problem-solving, negotiation, and anything where the “right” answer depends on context that’s hard to codify. These stay human.
The mistake companies make is either (a) assuming nothing can be automated because “our business is different” or (b) assuming everything can be automated because a vendor told them so. The truth is boring and specific: some of your processes are great automation candidates, some aren’t, and the only way to know which is which is to actually look.
$174,720
That’s the number. Not a projection. Not “what AI could do someday.” It’s what’s sitting in the payroll line of a typical 40-person professional services firm right now, every year, spent on work that doesn’t require human judgment.
Most companies will spend $200K hiring two more people this year instead of spending $40K to fix the workflows that are eating their existing team alive.
I get it. Hiring feels like doing something. Process work feels like homework.
The math doesn’t care about feelings.
