The boring AI wins

Every week someone shows me an AI demo that makes the room go quiet. An agent that researches a market, writes a strategy, books meetings and reports back. It is impressive. It is also, in most small companies, the wrong place to start.

The automations that actually survive inside a business are almost always boring. They take one repetitive task that a real person does every day, and they make it disappear. Nobody claps. Somebody just goes home thirty minutes earlier.

Why the impressive demo fails

A big autonomous system has a lot of places to break. It needs clean data, clear permissions, a decision about who checks its output, and someone who owns it when it misbehaves at 11 p.m. on a Friday. Large companies have people for each of those jobs. A ten-person company has the founder and maybe one operations person who is already overloaded.

So the ambitious project stalls. Not because the technology doesn’t work, but because the organisation around it isn’t ready. And after one stalled project, the team decides “AI isn’t for us,” which is the most expensive conclusion of all.

What a good first automation looks like

When I look at a small team’s week, I look for a task that has four properties:

• It happens often. Daily or several times a week. Rare tasks don’t pay back the setup time.
• The input is predictable. The same kind of email, form, file or order every time.
• A wrong answer is cheap. If the automation gets it wrong, someone notices and fixes it in a minute.
• One person feels the pain. That person becomes the owner and the first fan.

Typical winners: turning incoming order emails into rows in a sheet, drafting first replies to common customer questions, renaming and filing supplier documents, summarising a long thread into three bullet points before a call, pulling yesterday’s sales into one message every morning.

None of these will impress an investor. All of them will be used every single day.

The test I use

Before building anything, I ask one question: if this worked perfectly, who would notice tomorrow morning, and what would they stop doing?

If the answer is a specific person and a specific task, it’s a good candidate. If the answer is “the company would be more efficient in general,” it’s a strategy deck, not an automation.

Simple doesn’t mean small

The interesting thing is what happens after the first boring win. The team starts to trust the system. They begin to spot the next repetitive task on their own. Data starts flowing into one place because the first automation needed it there. Six months later, that ambitious agent from the demo is suddenly realistic, because the foundations exist.

Boring is not the ceiling. It’s the on-ramp.

Start with the task someone complains about every day. That complaint is your roadmap.