Course · 8 weeks
AI marketing academy: a free self-paced course
Eight weeks, eight skills, eight things you keep. No signup wall, no certificate, no fluff.
- 8 weekly modules
- 2-3 hours a week
- Free, no signup
AI marketing academy: a free self-paced course: what should you actually do?
This is a free eight-week course in using AI for marketing, written for people running their own marketing. Each week is one skill, one exercise and one output you keep: a context file, a research habit, a content system, an email sequence, an ad workflow, an analysis routine, an automation, and a measurement check. Budget two to three hours a week.
Last updated 3 August 2026
| Play | Effort | Cost |
|---|---|---|
| Week 1 — Build your context file | 90 minutes, once | Free |
| Week 2 — Research that cites its sources | 2 hours | Free |
| Week 3 — A content system, not content | 3 hours | Free |
| Week 4 — Email that sounds like a person | 2 hours | Free |
| Week 5 — Ads: variants and the honest test | 2 hours | Free to plan, ad spend optional |
| Week 6 — Analysis you can trust | 2 hours | Free |
| Week 7 — One automation that saves an hour a week | 3 hours | Free tier usually enough |
| Week 8 — Measure whether any of this worked | 90 minutes | Free |
Most AI marketing courses teach prompts. Prompts are the least durable part of this — the models change every few months and yesterday's clever prompt structure becomes unnecessary. What lasts is knowing which marketing jobs AI is good at, which it is bad at, and how to check the difference before you publish something wrong.
This course is built around that. Each week you learn one skill and produce one artefact you keep using: a context file, a research routine, a content system, and so on. By week eight you have a working setup rather than a folder of notes.
It assumes no technical background and no budget beyond a single $20 model subscription, and most weeks can be done on free tiers. Work through it in order — each week uses what the previous one produced.
- — Founders and owners doing their own marketing with no dedicated marketing hire
- — Marketers in a team of one to five who have used ChatGPT casually but not systematically
- — Anyone who has tried AI for marketing, got generic output, and assumed the tool was the problem
- — People with two to three hours a week for eight weeks, not a weekend intensive
Enterprise teams needing governance, procurement and legal review of AI usage, or anyone looking for a certificate to put on a CV. This teaches the work, not the credential.
The moves
Ranked, highest return first.
Work down the list. Each one names the first step so there's nothing to plan.
- 01
Week 1 — Build your context file
Generic output is almost always a context problem, not a model problem. A model that knows your product, customer, tone and last six months of results gives dramatically better answers than one starting from nothing.
First step: Write 600-800 words covering: what you sell, who buys it, why they buy, your three closest competitors, your pricing, your tone, and the three marketing things that have worked best so far. Save it and paste it at the start of every session, or store it in a project.
- Tools
- ChatGPT or Claude, a plain text file
- Effort
- 90 minutes, once
- Cost
- Free
- 02
Week 2 — Research that cites its sources
Models invent market sizes, competitor prices and search volumes with total confidence. Learning to separate a sourced answer from a plausible one is the single most valuable skill in this course.
First step: Pick one competitor. Research their pricing, positioning and customer complaints using a tool that shows sources, then open every source and check it. Note how many claims did not hold up — for most people it is one in five.
- Tools
- Perplexity, competitor sites, review platforms
- Effort
- 2 hours
- Cost
- Free
- 03
Week 3 — A content system, not content
One-off AI posts do nothing. A repeatable system — a source of topics, a brief format, a draft process, an editing pass that adds your own material — produces content that is worth publishing.
First step: List 20 questions your customers actually ask, from emails and calls. Turn the top three into briefs, draft each with your context file attached, then rewrite every paragraph that could have been written about any company in your industry.
- Tools
- Your inbox, ChatGPT or Claude
- Effort
- 3 hours
- Cost
- Free
- 04
Week 4 — Email that sounds like a person
Email has the highest return of any channel available to a small business, and it is the channel where AI output most obviously sounds wrong. Fixing that is a specific, learnable edit.
First step: Draft a four-email welcome sequence with the model, then delete every sentence that does not contain a fact, a number or a story. What remains is usually 40% of the draft and twice as good.
- Tools
- Your email platform, ChatGPT or Claude
- Effort
- 2 hours
- Cost
- Free
- 05
Week 5 — Ads: variants and the honest test
Generating 30 ad variants is trivial with AI; knowing that you cannot test 30 variants on a $20-a-day budget is the actual skill. This week is as much about statistics as writing.
First step: Write three genuinely different angles — not three rewordings — for one product. Work out how many clicks you need before a difference means anything, then run only as many variants as your traffic supports.
- Tools
- ChatGPT, your ad platform, a sample-size calculator
- Effort
- 2 hours
- Cost
- Free to plan, ad spend optional
- 06
Week 6 — Analysis you can trust
Uploading a spreadsheet and asking what it means is one of the highest-value AI uses in marketing, and one of the easiest to get quietly wrong when the model misreads a column.
First step: Export the last 90 days of traffic or sales data, ask for the three biggest changes and the likely causes, then verify each claim manually against the raw numbers before acting on any of it.
- Tools
- Your analytics export, ChatGPT or Claude
- Effort
- 2 hours
- Cost
- Free
- 07
Week 7 — One automation that saves an hour a week
One reliable automation beats five fragile ones. The aim is a single job that runs without you and does not need checking.
First step: Pick the most repetitive weekly marketing job you do — new-lead follow-up, review replies, repurposing one post into three formats — and automate only that. Run it manually for two weeks first to be sure the process is right.
- Tools
- Zapier or Make, your existing tools
- Effort
- 3 hours
- Cost
- Free tier usually enough
- 08
Week 8 — Measure whether any of this worked
AI makes output volume go up almost immediately, which feels like progress and often is not. The final skill is checking the numbers that pay you rather than the ones that flatter you.
First step: Compare the eight weeks against the eight before: leads, conversion rate, revenue, and hours spent. Keep the two activities that moved a real number and stop the rest.
- Tools
- Analytics, your sales records
- Effort
- 90 minutes
- Cost
- Free
Sequence
What to do first, week by week.
Foundations
Build the context file and learn to verify what the model tells you. Nothing else in the course works properly without these two. Expect week two to be uncomfortable — most people discover their previous AI research contained errors they published.
Content and email
Produce and publish real work: three pieces of content and one email sequence. Publishing matters more than polishing here; you need the feedback to know whether your editing pass is strong enough.
Paid and analysis
Move from producing to deciding. These two weeks teach the judgement layer — how many variants your traffic supports, and how to check an AI analysis against the raw data before you act on it.
Systems and proof
Automate one job and audit the whole eight weeks against revenue. Most people find two of the eight activities produced everything and drop the other six, which is the correct outcome.
Avoid
Where this usually goes wrong.
Treating the model as a source of facts
It is a source of drafts and structure. Every number, date, price and statistic needs a link you have opened. This one habit separates people who use AI well from people who publish confident errors.
Doing all eight weeks in a weekend
The spacing exists because you need real results between modules. Without a fortnight of published work in front of you, week eight has nothing to measure and the course becomes theory.
Publishing first drafts
First drafts are competent and forgettable, which is the worst combination for search and for readers. The editing pass where you add your own data and opinions is where the value is created.
Buying tools before week seven
Almost every module runs on free tiers plus one $20 subscription. Buying a stack in week one guarantees you pay for tools before you know which jobs you actually repeat.
Scaling output before checking direction
AI makes it easy to produce ten times more marketing aimed at the wrong customer. Weeks two and eight exist to catch that, and they are the two people most often skip.
Questions
Common questions.
Is this AI marketing academy really free?
Yes, the whole course is on this page with no signup wall and no certificate to buy. You may want one paid model subscription at about $20 a month, but every module can be completed on free tiers.
How long does the course take?
Eight weeks at two to three hours a week, so roughly 20 hours total. The pacing matters — each module uses results from the previous one, so compressing it removes most of the value.
Do I need technical skills or coding?
No. The only remotely technical module is week seven's automation, which uses a no-code tool and a single workflow. Everything else is writing, reading and checking numbers.
Which AI tool should I use for the course?
ChatGPT or Claude for the drafting and analysis modules, and Perplexity for anything involving research, because it shows sources. Do not switch tools mid-course; consistency makes your context file more useful.
What will I actually have at the end?
A context file the model reuses, a verified research routine, three published pieces of content, an email sequence, an ad testing framework, an analysis habit, one live automation, and a before-and-after read on your own numbers.
More playbooks
Other plans worth reading.
How to build an AI marketing strategy
ReadPlaybook · ContentHow to use AI for content marketing
ReadPlaybook · Digital marketingAI digital marketing: what to run, channel by channel
ReadPlaybook · ChatGPTChatGPT for marketing: the workflows that hold up
ReadPlaybook · ContentAI-powered content creation without publishing filler
ReadPillar · AI marketingAI marketing: the complete, honest guide
ReadElsewhere on gtm.help
Keep exploring.
Free marketing tools
Calculators and generators that run without a signup.
BrowseTool guides
What to use, what it costs, when it earns its keep.
BrowseTool comparisons
Head-to-head picks for the calls people get stuck on.
BrowseMarketing checklists
Task lists by business type and competitive situation.
BrowseStraight answers
The growth questions founders ask, answered with numbers.
BrowseGet started
Get this ranked for your business.
Paste your URL and you'll have a plan built from your site and competitors in about a minute.