"We're Not Big Enough for AI Yet" — Why $20M Companies Have It Backwards

"We're not big enough for AI yet."

This is one of the most common things I hear from $15M to $50M companies, and it's backwards.

Who actually benefits most from AI automation

The Fortune 500 has a different relationship with AI than you do. They have teams of analysts to throw at manual work. A report that takes three people four hours to build is a rounding error when you have 200 people in the finance function. It still gets done. Someone's job description even includes it.

You don't have that luxury. In a 50-person company, when three people spend four hours on a weekly report, that's meaningful overhead. When one person is the only one who knows how to pull a particular data export, that's a business continuity risk. When your ops manager spends half their time on tasks that follow a script, that's not a small problem — that's half a headcount.

That's the case for automating, not against it.

The size where AI moves the needle

Mid-sized companies sit in a specific sweet spot:

• Big enough to have real repetitive workflows — the kind that recur weekly, follow the same steps, and involve moving data between systems
• Small enough that 10 hours a week back is a meaningful chunk of someone's job, not a rounding error
• Lean enough that the people doing the manual work are often the same people who would be more valuable doing something else

A 500-person company automating a workflow saves labour cost. A 50-person company automating the same workflow frees up a person who was already stretched. The second effect is often larger, and it shows up faster.

The objection behind the objection

When mid-sized companies say they're "not big enough for AI yet," what they usually mean is one of a few things:

• They've seen what AI looks like when enterprise companies do it — $2M transformation programmes, dedicated AI teams, multi-year roadmaps — and concluded it's not for them
• They've tried something and it didn't work, so they've categorised the whole space as aspirational
• They're worried about disrupting the workflows that currently work, however imperfectly

None of those objections are really about size. They're about scope. The answer to all three is the same: start with one workflow, prove the ROI, and expand.

What "starting small" actually means in practice

It does not mean a 6-month pilot with a steering committee. It means picking the workflow that clears the four-criteria filter — weekly, consistent, data movement, dreaded — and automating the highest-friction step within it. Usually that's a three-to-six week project for a competent team working with clean data.

The metric at the end is simple: hours per week before vs. hours per week after. If the automation saves the equivalent of one full day per week for one person, it typically pays for itself within six months.

That's the conversation. Not "AI transformation." Not "enterprise AI strategy." One workflow, one measurable result.

If you want to understand where your company sits against industry benchmarks before starting, the AI Readiness Assessment.

maps your data maturity, workflow candidates, and quick-win opportunities in a structured format.