The best first job to give AI is not a big project. It is the small, dull, repetitive task that made you sigh this morning. Most owners wait for the perfect moment to "do AI properly" and never find it, while the real wins sit in the boring jobs they already do every week. Start there, with one annoyance, and you will get value faster than any grand plan would deliver.
Why the big AI project is the wrong place to begin
Say the word "AI" to most business owners and they picture something enormous. A system to build. A consultant to hire. Months of work and a budget to match. So they file it under "one day when we have time," and that day never comes, because there is never time for the big version.
Meanwhile the genuinely useful work is hiding in plain sight. Reformatting the same report every Monday. Chasing the same missing information. Copying numbers out of one system and into another. None of it is glamorous. All of it is exactly what AI is good at, and it is the easiest thing to hand over and the fastest to pay you back.
The real barrier is not the technology
Here is the part worth sitting with. The person who helped build one of these tools has said the biggest blocker to getting value from AI is not what the tool can or cannot do. It is that people do not think to ask.
We have been quietly trained out of asking. Years of "the computer can't do that" taught us to route around the annoying bits by hand and never question it. Children, interestingly, are often better at this than adults, because nobody has taught them the limits yet, so they simply ask for whatever they want and are surprised less often than you would think.
The unlock is not a new skill. It is a habit: when you catch yourself doing something repetitive and non-creative, pause and ask whether AI could just do it. Then go one level up. If AI can draft the reply, can it draft the reply from the details already sitting in your inbox. That "one level up, then repeat" move is where most of the real gains come from, and it is closely related to building automation that adapts as your process changes rather than breaking the first time something moves.
How to find your own starting point
You do not need a strategy document. You need a week and a notepad.
For one working week, keep a note on your desk. Every single time a task makes you sigh, something fiddly, repetitive, and not creative, write it down. Do not judge it, just capture it. By Friday you will have a short list, and it is the most honest map of your own time you are likely to make.
Then take the most annoying item on that list and ask whether AI could handle it, or at least most of it. That one task is your starting point. Not the biggest opportunity, not the most impressive, just the one that wears you down most for the least reason. Get that working, feel what it is like to have it off your plate, and the next candidate will suggest itself.
Some of those tasks will turn out to need a proper repeatable workflow rather than a one-off prompt, and knowing the difference between a simple workflow and a full agent helps you scope it sensibly. But you do not need to know that on day one. You just need to start with the sigh.
What we actually do at the start
When we begin with a client at Handiwork, the first work is rarely clever. It is walking through the week together and spotting the sighs, the little repetitive jobs nobody thought to question, and quietly handing the right ones to AI. It is unglamorous and it works, and it is a far better first step than any ten-tool rollout. If you would like a hand doing that mapping for your own business, that is what our practical use-case work is built around.
Ready to find out where you stand?
If you are not sure which of your weekly jobs are worth handing over first, our free AI Readiness Check helps you spot them. No cost, no pitch.
Frequently asked questions
What is the best first task to automate with AI?
The small, repetitive, non-creative job you already do every week and quietly resent, such as reformatting a report, chasing information, or moving data between systems. It is easy to hand over and quick to pay back.
Do I need a strategy before I start using AI?
No. Most businesses get further by starting with one annoying task than by writing a plan. Build the habit of handing small jobs to AI first; the strategy becomes obvious once you have felt where it helps.
How do I find the right tasks in my own business?
Keep a note for one week and write down every task that makes you sigh. That list is your shortlist. Start with the most annoying one.
What if AI gets the task wrong?
Treat it like a new team member, not a broken machine. Show it what you expected and where it went off track. Most of the time the fix is clearer instructions or cleaner input, not a different tool.
Ready to find out where you stand?
Take the free five-minute AI Readiness Check. There is no pitch at the end of it.
Take the AI Readiness Check
Sources
- How I AI: how the engineer behind Claude Cowork actually uses Claude Cowork — Lenny's Newsletter (Felix Rieseberg interview), https://www.lennysnewsletter.com/p/how-i-ai-how-the-engineer-behind
