Making Money With AI Starts With One Sum

Everybody wants to make money with AI. The tools are cheap, fast and getting better every month. You can build a website, an app or a chatbot in an afternoon. But building is no longer the hard part. Getting paid is. Last Tuesday, I wrote about the teams at Future Caribbean's Demo Day in Barbados and the Caribbean AI Summit hackathon in Puerto Rico. The next step for each of those teams is also your next step: find someone who will pay, and make sure you keep some of the money. So before you spend a dollar on an AI idea, answer these five questions.

1. Who has this problem badly enough to pay?

"Everybody" is not an answer. Pick a specific person: a clinic owner, a hotel manager, a real estate agent or a small shop owner. Ask them how they handle the problem now and what it costs them. If they shrug, keep looking. Researching demand, customer numbers, existing alternatives and pricing helps you understand the opportunity before you spend money on it.

2. Who signs the cheque?

The person who uses your product may not be the person who pays for it. A receptionist might use a booking tool, but the owner approves the spend. A supervisor might use a scheduling tool, but management controls the budget. Sell to the person with the budget. Your offer has to make sense to them, not only to the person using the product.

3. Will they pay your price before you build?

Liking an idea costs nothing. Put a clear offer in front of them: what they get, how long it takes and what it costs. One small paid trial tells you more than a hundred compliments. You can often deliver the first version by hand, with AI helping you, before you build anything. Describe any manual work honestly. Do not present an unfinished service as a fully automated product.

4. What does it cost you to deliver?

Count your tools, subscriptions, setup, fixing mistakes, support and your time. Your time is not free. If you are using free credits now, set your price as if they ran out tomorrow. A price above your software bill is not necessarily a profitable price.

5. What does it cost you to get each customer?

Add up your ads, content, events, meetings and follow-up calls. Divide that by the number of people who actually paid. That number is your customer acquisition cost. It includes more than advertising, and a lead, a follower or an unpaid trial is not a paying customer. Likes, followers and "interested" do not count. Only paying customers count.

The sum

Price, minus what it costs you to deliver, minus what it costs you to win the customer, equals what you keep before your running costs. Here is a made-up example in Barbados dollars: Item Amount You charge a small business to set up an AI booking assistant BDS$500 Cost to deliver it BDS$200 Cost to win that customer BDS$200 What you keep BDS$100 That BDS$500 sale leaves you BDS$100, and that is before rent, phone, internet and your accountant. For subscriptions, you should also track how long customers stay and how much profit they generate over time. Do not assume a first customer will keep paying forever.

A word on ads

If you run ads, someone will tell you about ROAS, or return on ad spend. Spend BDS$500 on ads, make BDS$1,500 in sales, and that is 3 to 1. It sounds great. But if delivering those sales costs you BDS$1,100, you lost BDS$100. Ad numbers measure sales, not profit. There is no useful ROAS target without understanding the costs and margins behind it.

What to do this week

Pick one idea. Name the person who pays. Talk to three of them. Write down your price. Do the sum. If the numbers work, build. If they do not, change the offer before you spend any more. Stopping an unworkable idea early is a good result too. It saves you money.

This is how you make money with AI

AI makes building fast. It will not find your customers, set your price or check your margins for you. That part is the business, and that is where the money gets made. This is what we teach in Beyond the Prompt and Build a Business with AI.

Build a Business with AI

Build a Business with AI is a one-day, in-person class on Saturday, 30 January 2027. You will work on the research, brand, website, content system and launch plan for your own business, using AI at every step. Price: US$199 Capacity: 50 seats Deposit: US$75 by 30 November 2026 secures your seat Reserve your seat: amplifiai.co/build

Janelle Germain

Founder & Chief Executive Officer Amplifi AI