The $768,000 Problem KC Businesses Don't Know They Have

The Number That Stopped Me Cold

$768,000. Per year. Twenty employees.

I've been in enterprise software and consulting for 37 years. I've sat across the table from hundreds of business owners who were convinced they had a cost problem, a people problem, a market problem. Almost none of them knew they had a process problem. And almost none of them knew what that process problem was actually costing them.

So when I ran the math on what manual administrative work costs a typical 20-person Kansas City business — using Bureau of Labor Statistics compensation data from late 2025 and some straightforward arithmetic — that number hit me differently than I expected. Employees spend 25–40% of their workweek on repetitive tasks that, in a meaningful percentage of cases, a governed AI agent could handle. At an average total compensation of $46.15 an hour, that's not a rounding error. That's $768,000 a year sitting in the drain.

Most KC business owners I talk to would shrug at that. "That's not real money." Except it is. It's just invisible. It's hiding in the 16 hours a week your operations manager spends re-keying data between systems, the invoices that go out late because someone had to chase down a number in a spreadsheet, the [27% of calls that go unanswered](/blog/why-27-percent-of-calls-go-unanswered) because nobody built the routing, the compliance task that got missed because it lived in a tab nobody opened.

The money isn't missing. It's just buried.

"We're Too Small for AI" Is the Wrong Frame

I hear this constantly. And I get it — "AI" conjures images of multi-million-dollar implementations, dedicated data science teams, 18-month rollouts. The enterprise software world spent decades making sure small businesses felt that way. It was good for the enterprise software world.

But that's not what we're talking about anymore.

The real problem isn't whether AI works at your scale. It's whether it's governed at your scale. And that distinction matters enormously for a company with 20 people and no IT department.

Here's the uncomfortable reality: 49% of SMBs cannot afford enterprise-grade custom software. Another 41% have no internal IT resources at all (these numbers come from recent SMB technology adoption research — they're not outliers, they're the norm). So businesses fill the gaps. The average SMB cobbles together nine cloud applications and holds the whole thing together with spreadsheets. Office workers perform over 1,000 copy-paste actions weekly. And 88–94% of business spreadsheets contain errors — one study put it at 94% of spreadsheets used in actual business decision-making containing material errors. That's not a productivity problem. That's a liability problem wearing a productivity mask.

When those businesses do try to bring in automation — no-code platforms, workflow tools, point solutions — Gartner found that 43% of those citizen developer initiatives were scaled back, paused, or shut down within three years. The leading cause? No governance framework. Sixty-one percent of failures traced back to that single gap.

So the "we're too small for AI" objection is actually backwards. You're too small to absorb the cost of ungoverned AI. That's a different problem with a different solution.

The Lawsuit You Don't Know Is Coming

Let me shift gears because there's a second cost buried in this that most KC business owners haven't priced in yet.

ADA website accessibility lawsuits in Missouri are up more than 35% year over year. These are cases filed against businesses — often small ones — whose websites don't meet accessibility standards under the Americans with Disabilities Act. The suits are frequently filed by serial plaintiffs working with specialized law firms, and the targets are chosen specifically because small businesses are less likely to fight back. They settle. Quickly. Quietly.

The economics here are brutal. Businesses with annual revenues of $10 million or less make up only 20% of commercial revenue but absorb 48% of commercial tort costs — roughly $160 billion nationally as of 2021, according to research from the Brattle Group commissioned by the U.S. Chamber of Commerce's Institute for Legal Reform. Normalized by revenue, small firms pay approximately $35 in tort costs per thousand dollars earned. Large corporations pay less than $5 for the same unit. Seven-to-one disparity.

A single ADA lawsuit, even a weak one, can run $15,000–$25,000 just to resolve without going to trial. For a 20-person company in Overland Park or Lenexa or the Crossroads, that's not an abstract risk. That's a real cash hit that lands in a bad quarter and doesn't come with a warning.

This is exactly where governed AI matters in a way that most vendors won't tell you, because most vendors aren't thinking about it. AI agents that touch customer-facing systems — your website, your intake forms, your scheduling — need to operate inside compliance guardrails. Not as a feature. As a requirement.

The Governance Gap Nobody Talks About

Here's where I'm going to push back on the AI industry a little, including the part of it I work in.

There's a lot of noise right now about AI agents — automated workflows, multi-step reasoning, systems that can handle complex tasks without human intervention. Most of it is genuinely impressive. Some of it is genuinely useful for SMBs. But the conversation almost always skips the part where someone asks: what happens when the agent does something wrong?

Because agents will do something wrong. Not catastrophically, usually. But they'll misclassify a customer record. They'll send a communication that violates a workflow rule. They'll act on stale data. And in a 20-person company with no IT department, nobody will notice until the damage is done.

I mapped out exactly where those operational leaks happen in [Four Streams Every Business Leaks](/blog/four-streams-every-business-leaks). Governed AI means the agent operates inside defined boundaries — boundaries that reflect your specific workflows, your compliance obligations, your industry context. It means there's an audit trail. It means a human can see what the agent did and why. It means the system fails safely rather than silently.

This is not a luxury feature for enterprise clients. For a Kansas City SMB, it's the difference between AI that reduces your operational risk and AI that creates new categories of it.

What This Actually Looks Like in Practice

I want to be concrete because I think the abstraction is part of why KC businesses aren't moving on this.

A 20-person professional services firm is spending maybe 500 hours a month on administrative work — scheduling, reporting, document handling, client follow-up, compliance tracking. That's not a guess; it's what the research bears out when you apply the 36% figure for entrepreneur time on administrative tasks and extend it across a small staff. Five hundred hours a month at fully loaded compensation rates is north of $60,000 a month. That's the $768,000 number annualized.

A governed AI agent framework — one that's actually designed for the scale and risk tolerance of a small business — can systematically reduce that burden without requiring an IT department to babysit it. The governance layer means the agents know what they're authorized to do, log what they've done, and escalate when something falls outside their scope.

That's not magic. That's just engineering done right for the actual customer.

The Position I'll Stand Behind

Kansas City has a genuinely strong SMB ecosystem. There are real businesses here doing real work — in construction, professional services, healthcare, logistics, creative industries — and they're competing against each other and against national players on thin margins. They don't have slack in the system.

The $768,000 isn't recoverable all at once, and I'm not going to pretend otherwise. But 20–30% of it is recoverable in the first year with the right approach. That's meaningful money. That's a hire, or a marketing budget, or a cushion against the next bad quarter.

What I've seen in 37 years is that the businesses that survive hard markets are the ones that got serious about their operational costs before the pressure hit. Not after.

The window for getting ahead of this — before competitors in your market move, before a compliance exposure becomes a lawsuit, before you've spent another year paying people to do work machines should be doing — that window is open right now.

It won't stay open.

If you want to see how much of that $768,000 is recoverable in your operation, [schedule a conversation](/schedule). We'll run the numbers together.