Why Is This Still Manual?

I ask this question constantly. My team hears it so often it's become a reflex.

Someone walks in with a process. A workflow. A thing they've been doing by hand. And before they get through the explanation, I'm already asking: why is this still manual?

Not to be difficult. Because every manual process that could be automated is a tax. At 10 people, it's manageable. At 35, it compounds. At 200, it becomes a structural ceiling on what the company can do.

The Belief Problem

The hardest part of pushing automation isn't the technology. It's the beliefs people hold about manual work.

I've heard every version of them. "This process is too complex to automate." "We need a human in the loop for this." "We tried once and it didn't work." "It's only 20 minutes a day."

None of these are wrong on their face. Some of them are even true. But they're all the same thing: a reason to stop asking. And once you stop asking, the process becomes permanent.

Here are the beliefs I push back on most:

What This Looks Like in Practice

At LlamaLab, we process thousands of medical record requests. Early on, parts of that process were manual. Status checks. Follow-up calls. Routing requests to the right provider channel. Flagging delays.

None of these felt like big problems at the time. Each one was maybe 15-20 minutes. Manageable.

But I kept asking the question. Why is this still manual? And the answer was always the same: because we hadn't built the automated version yet. Not because a human needed to do it.

So we built it. Status checks became automated tracking. Follow-ups became triggered workflows. Routing became rule-based logic. Flagging became real-time alerts.

The result: we handle volume today that would require three times the headcount at a company that hadn't asked the question. Not because our team is smarter. Because they're not carrying work that shouldn't require a human.

The Disruption Is Never the Product

Here's the thing most people miss when they study companies that changed industries: the story gets told as a product story. It's almost never a product story.

The disruption is the operating model. A different ratio of automation to manual effort. A different set of assumptions about what requires a human and what doesn't.

McKinsey's November 2025 research found that 57% of U.S. work hours are automatable with technology that already exists. That's not a future projection. That's today. Most companies are operating with more than half their work hours on tasks a machine could handle.

The incumbents we compete with have more resources. More brand recognition. More years in the market. The only way to compete is to operate differently. And operating differently starts with refusing to accept the same assumptions they built their operations on.

Palantir's CEO said it plainly: "We will grow 10x with fewer employees than we have today." SaaStr went from 20+ humans to 3 humans plus 20 AI agents and tripled their content output. These aren't outliers. They're the model.

The Scaling Trap

At 5 people, everyone questions everything. There's no room for waste. When something is manual, someone notices immediately because they're the one doing it.

At 35, it changes. Process accumulates. People build workflows that work for them. Habits form. The instinct to challenge gets replaced, slowly, by the instinct to protect.

This is the trap. And most companies fall into it without realizing it's happening.

The companies that stay fast as they scale are the ones that actively fight this. Not by being chaotic. By being deliberate about which processes deserve to exist and which ones are just inertia.

I make it part of how we operate. Every process is a hypothesis. Every manual step is a candidate for elimination. The question isn't "is this working?" The question is "should this still be manual?"

The Human Side

The hardest version of this question isn't about tasks. It's about people.

As a company grows, people build expertise in the manual version of work. They get good at it. They're valuable because of it. And then the automation comes and the question becomes: what do they do now?

This is the real friction. Not the technology. The human side of transition.

The answer, every time we've faced it: the people who were doing the manual work are the best people to own the automated version. They know where the edge cases are. They know what breaks. They know what the system can't handle yet.

Automation doesn't replace them. It changes what they do. The best operators own that transition.

The Bottom Line

Building a startup is a constant argument with the way things have always been done.

The argument isn't always comfortable. It's not always fast. Sometimes the manual process exists for a real reason, and you learn that by questioning it. But the companies that win are the ones that keep asking. That treat every manual step as a hypothesis to test, not a fact to accept.

The question isn't complicated. It's just three words.

Why is this manual?

If you don't have a good answer, you have your next project.

Sources: McKinsey Global Institute, Nov 2025, Wedbush Agentic AI Report, Jan 2026, FutureTask Startup Automation Research 2026, SaaStr on AI Agents, Palantir CEO on 10x Growth.