Every owner I talk to wants AI in the business. Almost none of them can name the actual job. They know they are behind, they feel the pressure, and then the conversation stalls the second I ask what AI should do on Monday morning.
So they buy a tool. They poke at it for a week. It sits there, doing nothing, while the subscription renews.
The AI Opportunity Map fixes the naming problem before you spend a cent on software. It reads how your week actually goes and hands you the five jobs where AI pays off first, ranked, with numbers. Ten questions. About five minutes.
The problem is not the tool. It is the target.
You have seen the demos. A chatbot writes an email. An agent books a meeting. It looks impressive for thirty seconds and does nothing for you by Friday, because none of it was pointed at work that costs you real hours.
The number of jobs AI can run inside a business is large. Quoting. Invoicing. The follow-up that never gets sent. The weekly report someone rebuilds by hand every Monday. Onboarding a new hire. Chasing overdue payments. Once you count the small repeated jobs nobody bothers to log, the list runs into the hundreds.
That is the trap. Too many candidates and no order. When everything qualifies, you freeze, or you grab whatever is trending and it saves you nothing.
You do not need more options. You need the five that pay off first, in the right sequence.
What is the AI Opportunity Map?
The AI Opportunity Map is a free assessment that looks at how your week is spent and gives you the five highest-impact jobs AI could take off your plate, ranked by time and money saved. It is scored for your business, not for AI in general. You answer, it ranks, you get a plan you can act on the same day.
That last part is the whole point. Most AI content tells you what is possible. This tells you what to do first.
Guessing your way in versus starting from the map
Here is the difference between the way most owners approach AI and the way the map approaches it.
| Guessing your way in | Starting from the map | |
|---|---|---|
| Where you begin | The tool everyone is posting about | The job that costs you the most |
| The order | Whatever looks shiny | Ranked by hours and money saved |
| First result | A tool that sits unused | Hours cleared off one real job |
| Time to know your move | Weeks of trial and error | About five minutes |
Neither column is magic. The right one just refuses to let you start from the wrong end.
How it works
Three steps. None of them need prep.
Step one. Answer ten questions. Mostly taps. They ask how your team spends a normal week: the jobs you repeat, who does them, roughly how many hours each one eats. No essays. You can do it on your phone between two other things.
Step two. Add your context, or skip it. If you have SOPs, process notes, or a spreadsheet of how things run, hand them over for a sharper read. If you do not have anything ready, skip it. The map still works on the ten answers alone.
Step three. Get your top five. Your answers are matched against real installs in $1M to $30M businesses, and you get five ranked opportunities with numbers attached: where an AI agent saves the most, and roughly how much. The results show on screen and land in your inbox as a PDF you can forward to whoever signs off on spend.
Done. You walked in wanting "AI, somehow." You walk out with five named jobs in priority order.
Get a sharper map: write your process note first
The map reads your week from ten answers. If you want it to read deeper, spend ten minutes first and turn your own head into a one-page note you can upload. This prompt runs in Claude or Codex and does the interview for you.
Copy: The Process Note Prompt
Copy this.
You are helping me prep for a short business assessment. Interview me one question at a time about how my team spends a normal week. Cover: the jobs we repeat every week, who does each one, roughly how many hours it takes, and where work stalls waiting on a person. Ask eight questions, one at a time. Wait for my answer before the next one. Then write a one-page process note. Plain language, no jargon, grouped under [SALES], [DELIVERY], [FINANCE], [ADMIN].
Answer it honestly. Upload the note when the map asks for context. You get a read based on how your business runs, not how you wish it ran.
Why five and not the whole list
Because a list of fifty things you could do is a way to do nothing. The point of ranking is to kill the paralysis. The top job clears the most hours for the least mess, so you have somewhere obvious to start and no excuse to stall.
Five is enough to see the shape of a plan and short enough to actually run.
What you do with the five
A ranked list is only worth something if you move on it. Here is the discipline that makes it pay.
Pick one. Not five. Start with the top job and finish it before you touch number two. Use this prompt to force the choice when you are tempted to spread yourself across all of them.
Copy: The First-Move Prompt
Copy this.
Here are the top five AI opportunities a map flagged for my business:
[PASTE YOUR TOP FIVE].
Pick the ONE I should build first.
Judge on three things: hours saved per week, how many people it touches,
and how little it depends on messy or missing data.
Give me the pick in one sentence.
Then give me the first three steps to stand it up this month.
Do not pick more than one.
The first picks tend to look boring, and boring is the tell that they are worth it. The follow-up that never gets sent, now sent every time. The weekly report someone rebuilds by hand, now built while they sleep. The quote that takes an hour, now drafted in minutes for a person to check and send. Dull jobs, real hours. That is where the money is, not in the demo that writes poems.
Honest: what the map will not do
You should know where it stops before you lean on it. Here are the real ones.
- It tells you what, not how. The map names the five jobs and the payoff. Building the agents is separate work. The map is the plan, not the install.
- It is only as honest as your answers. Round your hours up to look busy and you get a fantasy map. Give it the real, unglamorous numbers.
- The numbers are estimates, not a quote. They come from patterns across real installs, not from an audit of your books. Read them as a ranking, not an invoice.
- It is tuned for businesses roughly between $1M and $30M. If you are pre-revenue or running a solo side project, the ranking has less to bite on. You still get signal, just less of it.
- If you already know your one bottleneck cold, skip the map and go build it. This is for the owner with ten candidates and no order, not the one who has clarity and no time.
If you do one thing this week
Take the ten questions. Five minutes, on your phone, between two other things. Get your five jobs ranked and read the top one out loud.
Then build that one, or hand it to someone who can. Not all five. One. If it clears three hours a week, that is roughly a hundred and fifty hours a year off a single job, and you stop guessing where AI fits for good.
If you hate it, you spent five minutes. That is the whole downside.