You have probably had some version of the same conversation three times this month. Someone mentions AI. You nod. You feel that quiet unease that comes with knowing you should probably be doing something about this, but you are not sure what. So you open ChatGPT, stare at the blinking cursor, and type something like “what can you do?” Then you close the tab.
That is not a you problem. That is the wrong starting point.
The question everyone asks first is the wrong one
Most people approach AI the way they would browse a hardware store when they have no project in mind. There are thousands of tools on the shelves, each one capable of remarkable things, and none of it is useful until you know what you are trying to build. So you wander the aisles, pick up a few things, put them back, and leave empty-handed.
Asking “what can AI do?” puts you in browse mode. You end up reading articles, watching demos, collecting tabs, and feeling more behind than when you started. The vendors do not help — they are incentivized to show you everything, which means you leave with an impression of overwhelming possibility and no idea where to begin.
The question that actually moves things forward is this: What is the most time-consuming, low-value task in my week that I would be relieved to hand to someone else?
That is your starting point. Not the tool. The task.
Start with the work you do not want to be doing
Here is the pattern I see in almost every small business I work with. The founder is good at their actual work — the service they provide, the relationships they hold, the decisions only they can make. But around that core work sits a layer of administrative friction: writing the same kind of email again, summarizing a document before a call, turning rough notes into something structured, drafting a proposal that follows a format you have written a hundred times before.
This is the work that AI handles well right now. Not because it is magic, but because these tasks are largely about following a pattern with good language. A capable assistant who had seen enough of your emails could draft them for you. That is, roughly, what a well-prompted AI tool does.
So rather than starting with the tool and hoping an application falls out, start with your calendar or your to-do list from last week. Look for the tasks that took real time but did not require your specific expertise. The kind of thing where you thought, while doing it, “this feels like something a smart assistant could handle.” Circle those tasks. That is your list.
Then, and only then, ask which tool might help with each one.
A simple way to run your first real experiment
Pick one task from that list. One. The smaller, the better.
If you write a weekly update to your team, try drafting it with AI next week. If you spend twenty minutes turning meeting notes into a summary, hand that job over. If you spend an hour researching a topic before writing something, describe what you need and let the tool bring you a first pass.
You are not adopting AI. You are running one small test. The only thing you are measuring is whether the output is good enough to save you meaningful time, even after you edit it. If you spend forty minutes producing something that would have taken you an hour, that is a worthwhile result. If it takes longer to correct than to write yourself, try a different task or a different approach.
Most people find their first useful win within a week or two of working this way. Not because the technology is extraordinary, but because they finally stopped asking it to impress them and started asking it to help them.
You do not need to know how any of this works under the hood. You do not need to understand the technology. You need to know what you want done and to be willing to try once. The rest you can learn as you go, or hand to someone who already knows it.
What this looks like in practice
A client came to me recently overwhelmed by her inbox. She was a consultant, and every new inquiry required essentially the same email — warm, professional, asking the right questions before a discovery call. She was writing each one from scratch. We spent thirty minutes building a simple template she could paste into ChatGPT with a few details about the sender, and the tool would produce a first draft in her voice. She edited for two minutes. Her response time went from days to hours, and the emails were better.
That was her starting point. One task. One test. Now she uses AI for three or four things in her business, each one chosen the same way: she noticed where her time was going and asked whether any of it could be handed over.
She did not start with a strategy. She started with an email.
If you are genuinely new to this, that is the permission I want to give you: you do not need a plan for AI in your business. You need a task. Find the thing in your week that drains time without requiring your best thinking, and try giving it to an AI tool once. See what comes back.
The question of which tool to use, and how to get better results from it, is a conversation worth having once you have your first experiment in mind.
