What AI Actually Is (and Isn't) for Marketers
Cut through the hype and understand what large language models do, where they excel, and where they fall short. You'll leave with a clear mental model that makes every future AI interaction more effective.
In practice: Campaign briefs: 3 hours → 20 minutes
Your version of this lesson adapts to your role. After the 3-minute assessment, examples, scenarios, and exercises are tailored specifically to your job function and experience level.
Personalise →The One Concept That Changes Everything
AI tools like ChatGPT, Claude, and Gemini are large language models (LLMs). They predict the next most-likely word based on patterns learned from billions of documents. That one fact explains both their remarkable capabilities and their key limitations.
What LLMs Are Genuinely Great At
- ■Generating first drafts fast. An LLM can produce a serviceable email subject line, blog intro, or social caption in seconds. It won't be perfect, but it eliminates the blank-page problem.
- ■Rewriting and reformatting. Paste in a long report and ask for a three-bullet executive summary. The model compresses and restructures fluently.
- ■Brainstorming at scale. Ask for 20 campaign angles and you'll get 20. Filter from abundance rather than straining for ideas.
- ■Tone-shifting. Take formal copy and make it conversational, or vice versa, without rewriting from scratch.
What LLMs Are Not
- ■They are not search engines. They don't retrieve live data. Information has a training cutoff date.
- ■They are not fact-checkers. They'll confidently state incorrect statistics. Always verify numbers before publishing.
- ■They are not strategic thinkers. They reflect patterns in training data, not genuine market understanding. Your judgement is still the product.
The Marketing Implication
Think of AI as a very fast, very well-read junior copywriter who never gets tired but sometimes makes things up. Your job shifts from writing everything to directing, editing, and verifying. The creative instinct and brand knowledge stay with you.
How to Approach Your First Interactions
Start with low-stakes tasks: rewriting an existing email, generating five headline options, or summarising a competitor's homepage. These let you calibrate the tool's output quality before trusting it with higher-visibility work.
Key principle: The quality of your output depends almost entirely on the quality of your input. Vague prompts produce generic content. Specific prompts produce useful content. The rest of this curriculum teaches you to write specific prompts.
The prompt
Claude's response
Quick check
An AI tool confidently states that a competitor's market share grew 34% last quarter. What should you do?
Key Takeaways
- ■LLMs predict likely next words — they don't understand content the way humans do
- ■AI excels at drafting, rewriting, summarising, and brainstorming — not fact-checking
- ■Your role shifts from writer to director and editor, which is a more strategic position
- ■Always verify statistics, dates, and factual claims before publishing AI-generated content
- ■Prompt quality directly determines output quality — specificity is the core skill
Your challenge this week
Apply what you learned in a real task
Pick one piece of content you write every week — a report, email update, or social post. Use an AI tool for the first draft with a specific prompt. Note how much editing it needs versus writing from scratch.
Before you practise
Think about the last piece of content you wrote at work. Which part took longest — coming up with the structure, the first draft, or the editing? AI is strongest at the part that slowed you down most.
Ready to put it into practice?
Apply what you just learned with a hands-on exercise.