Home  /  AI Courses  /  Paid Ads on Autopilot

For business owners and marketers

Most people who try Facebook ads quit on day three, after one advert, and decide it does not work.

They were three days into something that takes three weeks. This course teaches the actual machine: how to use AI to test forty angles for the price of one, find the winner, and then defend it once you have it. Written for Uganda, in shillings, on a budget of 20,000 to 35,000 a day.

10Modules
37Lessons
1Free and complete
Before anything is sold to you

The promise is real. It only becomes true after two things are true.

Paid advertising genuinely can become a machine where you put one shilling in and three come out. People run that machine every day.

That machine only exists after you have found a winning creative and you have an offer that works. Before that point, ad spend is not a printing press. It is a search cost. You are paying money to discover something you do not yet know, and that is a completely different activity. It has to be budgeted, measured and emotionally handled differently.

Almost everybody who says ads did not work for them quit in the middle of that search, having tested one advert, and concluded the channel was broken. What actually happened is they paid for information and walked away before collecting it.

So what does AI actually do? It does not make the printing press print faster. It collapses the search. An agency tests six creatives a month because each one costs money and days. With Claude you test sixty. Then, once you have the winner, the same system watches it daily, spots fatigue before it costs you, and has the replacement ready.

Find faster. Defend longer. That is the whole thing. If anyone tells you AI makes ads profitable on day one, they are selling you something. This is the version that is actually true, and it is the version taught inside.

The arithmetic

Every decision in this course is a number, not an opinion.

What your budget can actually buy in signal. The cost per result that means scale. The one that means kill it today. Written down as thresholds, so you never have to guess in the middle of a losing week.

A row of shops and boda bodas on a street in Uganda

These are the businessesWritten for shops like these, on 20,000 to 35,000 shillings a day

Fit

Who this is for, and who it is not for.

For you if

You already sell something

  • You sell something, and you have sold it at least once without advertising.
  • You want to run your own acquisition instead of paying an agency you cannot audit.
  • Your budget is small, 20,000 to 35,000 a day, and you need it to teach you something rather than just disappear.
  • You would rather do the arithmetic before you do the creative.
Not for you if

You are looking for a way to make money without a product

  • There is no such thing in here, and pretending otherwise would waste your money.
  • You want to be told your offer is fine. Module 1 will probably tell you it is not.
  • You are not willing to look at your own numbers honestly.
  • You want results this week. The search takes about three weeks before it tells you anything trustworthy.
The curriculum

Ten modules. Thirty seven lessons. Each one an operation you perform.

Every lesson below is a single outcome with a prompt you copy, the steps to run it, what goes wrong, and one thing to practise on your own business. The full lesson titles are published here so you can read all of them before you decide.

Module 01 · 3 lessons

Whether you are allowed to spend yet

The only module that decides whether the other nine matter. By the end you can say, in numbers, whether your business can survive buying traffic, and if it cannot, exactly which number has to move first.

  • How to work out whether you can afford to buy traffic at all
  • How to find out whether your offer can survive paid traffic
  • How to choose what to optimise for when your budget is small
Lesson one of this module is free and complete
Module 02 · 4 lessons

Being able to see what your money did

A large share of "my ads do not work" is actually "my ads work and I cannot see it". Wrong numbers are worse than no numbers, because you act on them.

  • How to design a tracking setup that reports the truth
  • How to tell a broken campaign from a broken measurement
  • How to name things so your own data can answer questions
  • How to reconcile three platforms with one bank account
4 lessons
Module 03 · 5 lessons

Producing twenty ideas instead of one

You are not trying to make one beautiful advert. You are running a search across a space of ideas, cheaply. This module is also honest about where the quality floor of AI creative currently sits.

  • How to turn five angles into twenty genuinely different adverts
  • How to generate static images that do not look like adverts
  • How to produce video creative without a camera or a face
  • How to reject the bad half of a batch in under a minute per asset
  • How to build the log that becomes your real advantage
5 lessons
Module 04 · 5 lessons

Generate fifty, ship five

The generating is easy. The cutting is the skill. Angle, hook and copy are three different things, and most people running adverts cannot separate them.

  • How to generate fifty angles that are actually different
  • How to cut fifty angles to five without using taste
  • How to write fifteen hooks for one angle
  • How to get copy that sounds like you wrote it
  • How to audit your own copy for machine tells
5 lessons
Module 05 · 3 lessons

Launching without hurting yourself

Simple structures beat clever structures, and both beat whatever you were about to build. This module stops the two standard self inflicted wounds: fragmenting your data, and resetting the learning phase by fiddling.

  • How to build the simplest structure that can actually work
  • How to spend one budget across three platforms without wasting it
  • How to audit a campaign before you spend anything on it
3 lessons
Module 06 · 3 lessons

Knowing which part is broken

Every metric is a question about one specific part of the funnel. Learn which part and diagnosis stops being instinct. This module also tells you how much data a number needs before it means anything, which is the part usually left out and the part that saves the most money.

  • How to find the first broken link in your funnel
  • How to tell whether it is the creative, the offer, the audience or the page
  • How to run a weekly review that is decisive rather than fidgety
3 lessons
Module 07 · 3 lessons

Rules you write before the money is on the table

Your judgement is fine. Your judgement while watching your own money disappear is not. Every decision becomes a numeric threshold written in the cold, including the one almost no course includes: the point at which you stop making adverts and go back and fix the offer.

  • How to write the rule sheet you will run instead of your feelings
  • How to turn today's numbers into verdicts instead of anxiety
  • How to scale a winner without killing it
3 lessons
Module 08 · 4 lessons

Automate the reading, never the spending

Connected properly, Claude can hold your whole operating system in context at once: your unit economics, your decision rules, your creative log and your live data. This module also draws the line between what the machine decides and what you decide, and says exactly what goes wrong when that line is crossed.

  • How to get a daily brief that arrives before any dashboard
  • How to make creative supply stop being the bottleneck
  • How to decide what to automate and what must stay yours
  • How to catch an automation that has silently broken
4 lessons
Module 09 · 4 lessons

Where the profit actually lives

Front end acquisition buys you a customer at break even. Everything after that is the business. Raising what a customer is worth raises what you are allowed to pay for one, which is how you outbid everybody else for the same traffic.

  • How to work out what a customer is really worth to you
  • How to make a lead worth something after day one
  • How to design a backend that lets you outbid competitors
  • How to shorten the gap between paying and the first real win
4 lessons
Module 10 · 3 lessons

Nine parts, one loop, running weekly

The assembly. Two loops rather than one, and a plain statement of what the machine does not do, which is the part a course that wanted to sell more copies would leave out.

  • How to find the one thing actually holding your business back
  • How to run the whole machine in about six hours a week
  • How to tell yourself the truth about where the business is
3 lessons
All ten modules are written and ready. Every lesson carries the prompt to copy, the steps, the mistakes people make, and a worked example on a real Uganda business. You get access the moment your payment is confirmed, and every future addition is included at no extra cost.
Also included

The parts you keep and use every week.

37 prompts

Every prompt in the course, in one library you keep. Written to be pasted, not admired.

One page

The scale and kill rule sheet. Numeric thresholds you write before launch and run instead of your feelings.

Real numbers

A live case study running through every module. A Uganda only course business, real shillings, the same thin budget you have.

Why the case study matters

The biggest wins in paid advertising are usually not in the ad account.

Every module is worked through on the same live business. Same product, same currency, same mobile money, same thin budget, same platforms. You see the actual unit economics: a 94 percent contribution margin, a break even return on ad spend of 1.06, and a target cost per lead that is derived rather than guessed.

You also see the day eight numbers, where the blended cost per lead looked acceptable and the per advert breakdown showed that a large share of every shilling was being lost between the click and the page finishing loading. That finding was worth more than every creative decision on the table, and it came from a number almost nobody looks at.

Your judgement is fine. Your judgement while watching your own money disappear is not. That is why the thresholds get written down before the campaign goes live, not after.

Matthew NtamaraMSc Data Science

Enrol

One payment. Mobile money.

Pay by MTN or Airtel mobile money. You get a prompt on your phone, approve it, and your access arrives immediately.

Free first UGX 0

Read a complete lesson and use the rule sheet before you decide anything.

  • Lesson one of Module 1, in full, exactly as buyers get it
  • The scale and kill threshold sheet, one page
  • The break even and target cost worksheet, filled in with your own numbers
  • Delivered on the page, instantly. Nothing arrives later by email.
Get the free kit
All sales are final. That is why a complete lesson and the rule sheet are free on this site right now. Read the lesson first. If the arithmetic in it does not change how you think about your own business, do not buy the other nine modules.
Questions

Answered straight.

Will this make me money?

It will make you faster and cheaper at finding out whether your offer can be sold profitably. If it cannot, this course tells you that in Module 1 and saves you months. If it can, this is the fastest route to finding the creative that proves it. Anyone answering that question with a straight yes is not being straight with you.

My budget is only 20,000 a day. Is that enough?

Enough to learn, yes. Enough to learn fast, no. Module 1 shows you the exact arithmetic of what your budget can and cannot buy in signal, and how to work inside it by optimising for a cheaper event than a purchase. The case study runs on the same budget you have.

I do not want to be on camera.

Neither does the person in the case study. Module 3 covers faceless video, AI presenters and animation specifically because of that constraint, and is honest about which of them currently look convincing and which do not.

I have already tried ads and lost money.

Then you are the target reader. Module 6 will probably tell you which of the four things was actually broken, and it is usually not the one people assume.

Is AI ad creative not obviously fake?

Some of it is, and that kind should be thrown away. Module 3 is blunt about where AI creative fails, what the current tells are, and why rejecting a third to a half of what you generate is the correct number rather than a waste.

Do I need to be technical?

You need to be able to follow instructions and do arithmetic. The automation module goes further than that, and it is the one module you can skip entirely and still run the whole system by hand.

How is this different from the office course?

Different course, different problem, sold separately. AI for the Modern Office is about the work on your desk. This one is about buying customers. Owning one does not give you the other.

Start here

Read the free lesson first. It is the one that tells you whether to spend at all.

If the honest answer for your business is no, you will know inside twenty minutes, and I would rather you found that out for nothing. If the answer is yes, the other nine modules are waiting.

Checkout

Pay by mobile money.

You will get a prompt on your phone to approve the payment. If it times out or you decline it, nothing is charged and you can simply try again.