Most advice about getting started with AI is a list of tools. That is the wrong end of the problem.
Here is what we would tell you to do over thirty days if you were sitting across from us and you had decided this year is the year. It costs nothing until week three, and by day thirty you will have a number that tells you whether to keep going.
Week one: find out where the time goes
Do not buy anything. Do not sign up for anything. Watch.
For five working days, keep a running note on your phone. Every time you do something that made you think I do this constantly or this should not be my job, write one line. Do not analyze it. Just collect.
At the end of the week you will have somewhere between fifteen and forty lines, and about six of them will be the same thing said differently. Those six are your list.
Two rules for reading it:
- Ignore the things that are hard. Hard is usually judgment, and judgment is not what you are automating first.
- Circle the things that are repetitive and get skipped when you are busy. That combination — dull, frequent, and unreliable — is where the money is.
Week two: write one of them down
Take the single most annoying item on that list and describe how it actually happens today, step by step. What starts it, what happens next, who does each part, where it stalls.
Twenty minutes on paper. We wrote a whole piece on this exercise because it is the step everyone skips and it is the one that decides everything.
You are looking for one specific thing: the step that is supposed to happen every time and only happens when someone remembers.
By the end of week two you should be able to finish this sentence: When ______ happens, ______ should happen next, every time, and right now it happens about ______ percent of the time.
If you can say that out loud, you are further along than most businesses that have already spent money.
Week three: do it manually, badly
Now you can start, and you still are not buying anything.
For two weeks, do that missing step by hand, every time, on purpose. Set an alarm. Put it on a whiteboard. Whatever it takes to make the unreliable step reliable using nothing but human stubbornness.
This feels like a detour. It is the most valuable two weeks of the month, for three reasons.
You find out if it matters. Some steps you were sure were costing you money turn out not to move anything. Better to learn that now, free, than after a purchase.
You learn what the step actually involves. Doing it deliberately twenty times surfaces the exceptions. The cases where you would not send the reminder. The customer you would call instead. Those exceptions are exactly what has to be specified later, and you cannot know them from memory.
You get a baseline. Count what happens. Replies, callbacks, payments, whatever the step is supposed to produce. Write the number down.
That number is the most valuable thing you will produce this month. Everything after this gets measured against it, and it belongs to you rather than to a vendor's case study.
Week four: decide, then buy narrowly
Now you know three things you did not know on day one: what the step is worth, what it involves, and what it produces when it happens every time.
Now you can shop, and you are a completely different buyer than you were on day one. You are not asking what can AI do for me. You are asking what will make this specific thing happen without me, and what does that cost.
Buy the smallest thing that does that. Not a platform. Not a suite. The one thing.
Then keep counting. Same number, same way, for the next thirty days. If it does not beat your manual baseline, you have learned something real and you are out one month's subscription rather than a year and a rebuild.
What you will have on day thirty
- A list of the six things that eat your week
- One of them described precisely enough to hand to someone else
- A real number for what that step produces when it happens every time
- Either a small working system, or solid evidence that this particular thing was not worth automating
Any of those four is worth the month. Having all four puts you ahead of most businesses that have been "looking into AI" for a year.
The part people skip
Weeks one through three have no software in them, and that is why most people skip straight to week four.
We understand the impulse. Buying something feels like progress and watching yourself work does not. But every rescue project we get called into failed in the same place — someone bought a solution to a problem they had never actually described, and then discovered during setup that nobody in the building agreed on what the problem was.
The three unglamorous weeks are the whole thing. The purchase is the easy part, and it goes well or badly depending entirely on what you knew before you made it.