AI course: Why everyone needs to understand the new technologies

Micky Weis

15 years of experience in online marketing. Former CMO at, among others, Firtal Web A/S. Blogger about marketing and the things I’ve experienced along the way. Follow me on LinkedIn for daily updates.

Kopieret

I still remember the first time I showed ChatGPT to my mom.

She’s a schoolteacher, and her first reaction was: “Okay, my students can now cheat in German in 5 seconds.” It was funny—and a little scary. But it struck me. This seemed like much more than just another boring tech trend. It felt like a real everyday tool, like Google—a new place for information.

Since then, I’ve gone deep into AI. I’ve built my own Custom GPTs across topics like fitness, health, investing, and personal finance. I’ve experimented with no-code AI projects using tools like Lovable, automated processes, and tested a ridiculous amount of AI tools.

The point? I’ve made all the mistakes—so you don’t have to.

This post is my attempt to give you a complete, honest, and practical guide to AI—whether you’re a beginner or already deep into prompts and workflows.

You don’t need to take a course with me. You can absolutely learn this on your own: a good place to start is by reading this all the way through 😉

Why understanding AI is critical today

We’re in the middle of a technological revolution, and AI is no longer reserved for researchers, engineers, or Silicon Valley. It’s a practical reality in our everyday and professional lives—and it’s moving fast. Companies across the globe are implementing AI solutions at a pace we haven’t seen since Steve Jobs introduced the iPhone.

That means that no matter if you work in marketing, finance, HR, law, production, or something else entirely, AI will impact your job.

So a basic understanding of AI isn’t just a good idea—it’s a necessity.

And we’re not talking about building complex neural networks from scratch or understanding every line in a Python script. It’s about knowing what AI can do, what it can’t, how it’s being used—and how you can apply it yourself. Especially Large Language Models (LLMs) like ChatGPT, Claude, and Gemini now make it possible to use AI in your everyday life without being a technical expert.

But you need to know how to ask the right questions. How to structure information.

How to evaluate output. And how to think strategically with AI as your co-pilot. That’s what this guide is about.

In recent years, I’ve taught businesses, leadership teams, and specialists about AI—both theory and practice.

AI training for you: want to get started with AI in practice?

Strategy & consultation – get help implementing AI in your daily work or organization.

Automation & no-code – we build custom solutions using tools like Lovable, Make, and GPTs.

Workshops & talks – introduce AI hands-on to your team or network.

BOOK A MEETING

What I’ve learned is this: once you understand the basic principles and tools, AI is no longer a threat—it’s a gamechanger. The best students aren’t necessarily the most technically skilled, but the ones who are willing to experiment and keep learning.

Those who understand that the future’s value doesn’t lie in replacing people with machines, but in enhancing human potential with machine capacity.

But there are also pitfalls. AI can hallucinate, give incorrect answers, or sound convincing even when it’s wrong. That’s exactly why a basic understanding is so important. Not just of how to use it—but how it works, how to ask the right questions, and how to evaluate the results. That’s the difference between being a passive user and an active AI-literate professional. You’ll learn to spot when a model is guessing—and when it actually knows what it’s talking about.

This guide covers everything you need to know to build an effective, practical, and grounded understanding of AI and Large Language Models.

It won’t get technical just for the sake of it—it’ll be concrete and actionable. You’ll get cases, tools, step-by-step examples, and most importantly: a strategic framework for how you as an individual or business can begin using AI in your everyday life.

Whether you’re a freelancer, a team leader, a marketing professional, or just curious about tech’s possibilities—this guide gives you the overview and tools you need to get started.

Not in five years. Not next year. But now. Because AI is already here.

The only question is whether you’ll be one of the people who understands and uses it—or one of those getting left behind.

A bit dramatic? Maybe. But I actually think it’s true.

What is AI, and how does it actually work?

Artificial intelligence—or just AI—is a term many use, but few can explain clearly.

At its core, it’s about getting machines to mimic human intelligence: thinking, learning, understanding language, recognizing patterns, and making decisions. But AI isn’t one technology—it’s a collection of techniques, some simple (like rules and if-statements), others highly advanced (like neural networks).

AI is often confused with automation. Understanding the difference is crucial:

• Automation: When we program a machine to perform a specific task—over and over—without variation.
  Example: an email autoresponder that sends a standard reply when someone contacts support.
• AI: When we enable the machine to understand, learn, and improve over time.
  Example: an AI chatbot that analyzes your question, understands context, and gives a tailored response based on thousands of previous interactions.
• AGI (Artificial General Intelligence) is the theoretical next step, where AI not only handles specific tasks—but thinks and learns like a human across domains.

What makes modern AI so powerful is its ability to work with enormous amounts of data—and learn from it. Using statistical models and neural networks, AI can identify patterns humans would never notice. And generally, the more data it gets—the better it performs.

Three branches of AI you should know

• Machine learning: The system trains on data and learns patterns (most common today)
• Deep learning: A subfield of machine learning using advanced neural networks (like GPT-4)
• Natural language processing (NLP): Enables AI to understand and generate language—what you use when talking to ChatGPT

AI is already integrated into your daily apps—from Gmail’s smart replies and Netflix recommendations to voice control in your car. The difference is that now—through tools like ChatGPT—you can interact with AI directly, not just through a hidden interface. That changes everything.

Next, we’ll zoom in on the technology that brought AI to the masses: Large Language Models. Because they make it possible to write, think, plan, analyze, and brainstorm—with AI at your fingertips.

Introduction to large language models (LLMs)

If AI is the engine powering the technological future, then large language models – or LLMs – are the gearbox. They make it possible to interact with artificial intelligence using natural language. That’s precisely why models like GPT-4, Claude, Gemini, and Mistral have caused such disruption: they’ve democratized AI. You no longer need to be a developer, engineer, or data analyst to use advanced AI. You just need to write a sentence.

But first, let’s clarify what a large language model actually is. In short, an LLM is an AI model trained on massive amounts of text from the internet, books, articles, manuals, forums, documents, and more. Its job is not to search for information, but to predict the next word in a sentence based on the context you provide. It sounds simple. But when scaled to billions of parameters and decades of text, the result is surprisingly intelligent.

An LLM is not conscious. It doesn’t think. It has no opinions or intent. But because it has learned language structures and patterns at scale, it can answer complex questions, analyze texts, write articles, code software, and even build strategic plans – all from your prompt.

Examples of how LLMs are used

• Writing marketing copy, blog posts, and product descriptions
• Summarizing PDFs, long emails, and reports
• Enhancing customer service through automated language-aware bots
• Generating code, debugging scripts, or writing SQL queries
• Brainstorming ideas and concepts when you’re stuck
• Translating text and adapting tone of voice to the audience

How it works on a technical level

We won’t go too deep into data science, but what’s important to know is that LLMs are typically built on transformer architecture – a kind of neural network model well-suited for understanding context. It calculates the probability of each new word based on the previous ones and builds entire sentences from your prompt. The more precise and contextual your prompt, the more relevant the output.

Why LLMs are a gamechanger

You interact with the computer as if you’re speaking to a human – through language. And because it’s a language you already use every day, the learning curve is short. But the quality of the output is 100% dependent on the input. LLMs are mirrors – and you decide what they reflect.

The next section dives into what makes the difference between good and poor AI use: prompting. This is the key to using AI effectively – whether you’re a private user, specialist, or business.

Prompting – the most essential AI skill

If you think AI is just about pressing a button and getting magic, you’re mistaken.
AI is about asking the right questions. This is called prompting. And this is where the real value is unlocked.

What is a prompt?

A prompt is the input you give a large language model like ChatGPT, Claude, or Gemini. It’s your question, your instruction – your brief. The better you are at framing it, the better the output you’ll receive.

Think of it as a conversation with a sharp assistant: if you mumble, it guesses. If you’re clear and precise, it delivers.

Bad prompts – and why they fall short

• “Write a text about SEO.” – Too broad. Result: generic, meaningless content
• “What is marketing?” – Feels like a Wikipedia article
• “Make some cool content.” – What does “cool” mean? Who’s the audience?

Bad prompts are unclear, too broad, and lack context. The result? Boring and superficial output.

Good prompts – and why they work

• “Write a beginner-friendly blog post on SEO for B2B companies in Denmark. Use an informal tone and include concrete examples.”
• “Create a Facebook ad for a new protein bar. Target audience: men aged 25-40 interested in fitness. Tone should be humorous and sharp.”
• “Write 5 variations of this text for A/B testing with a focus on click-through rate.”

Good prompts are specific, targeted, and contextual. They tell the AI exactly what you want – and who it’s for.

Prompt strategies that work

Here are some methods I personally use that deliver reliable results:

1. The role-based approach

Have the AI take on a role:
“You are an experienced email marketing specialist. Write a newsletter for subscribers who haven’t engaged in 30 days.”

2. Frame + format + audience

Use this structure: “Write [content type] about [topic] for [audience] with [tone].”
It forces you to define your goal clearly.

3. Iterative prompting

Start simple and build from there:
“Give me an outline for a LinkedIn post about the future of marketing.”
(Then:) “Expand point 2 with more detail.”
(Then:) “Rewrite with a more edgy and concrete tone.”
Use AI as a sparring partner, not just a text generator.

4. Example-based prompting

Feed the model with examples:
“Here’s an ad that worked well. Write three variations in the same style.”
AI loves patterns – give it something to learn from.

Mini-guide: How to write a good prompt

• Start with purpose: What are you trying to achieve?
• Choose a format: Blog post, ad, video script, email?
• Define your audience: Who is it for?
• Specify tone and style: Should it be funny, serious, direct?
• Add context: Industry, product, past experience
• Iterate: Adjust and improve based on the output

Bonus: Prompt templates you can use now

• “You are a [role]. Write a [content type] about [topic] for [audience] with [tone].”
• “Brainstorm 10 ideas for a [format] on [topic], targeted at [audience].”
• “Rewrite this text to make it sound more [adjective]: [insert text]”

You won’t get good at AI by reading manuals. You’ll get good by learning to prompt properly.
It’s a skill – just like copywriting or sales. And it can be learned.

The sharper you are with prompting, the better your AI return. So: practice. Test. And find your go-to prompts.

AI tools: Getting started

You don’t need to be a developer to get started with AI. There are now hundreds of tools that make it easy to integrate AI into your daily work – whether you’re in marketing, content, video, sales, or something entirely different.

Here’s a breakdown of 10 powerful AI tools – with pros, cons, and what to watch out for.

1. ChatGPT (OpenAI)

Use for: Text generation, brainstorming, coding, customer support
Pros: Extremely versatile. GPT-4 Turbo is fast and accurate
Cons: Output can become generic without good prompts
Pro tip: Use “Custom GPTs” to build specialized assistants with custom instructions

2. Perplexity AI

Use for: Research and fact-checking with source citations
Pros: Fetches fresh data from the web. Fast and precise search
Cons: Not suitable for long content or creative tasks
Pro tip: Use the “Copilot” feature to guide your research with follow-up questions

3. Jasper

Use for: Copywriting and content for blogs, emails, and social media
Pros: Pre-built templates and tone-of-voice profiles
Cons: Subscription required. Less flexible than ChatGPT
Pro tip: Combine Jasper with SEO tools to write optimized content faster

4. Synthesia

Use for: AI-generated videos with avatars and voiceovers
Pros: Make videos without camera, studio, or editing
Cons: Voices may sound robotic depending on language
Pro tip: Use for onboarding, product demos, or international scaling

5. Runway

Use for: Video editing, AI videos, green screen effects
Pros: Powerful editing tools and text-to-video features
Cons: Requires some learning. Long render times for heavy tasks
Pro tip: Use “Gen-2” to create short AI videos from prompts

6. MidJourney

Use for: Visual content – illustrations, images, mockups
Pros: Exceptional image quality and creativity
Cons: Only available via Discord. Steep learning curve for beginners
Pro tip: Use clear style references in your prompts (e.g. “in the style of Pixar”)

7. Notion AI

Use for: Productivity, documentation, automatic summaries
Pros: Integrated into Notion. Great for internal notes and project planning
Cons: Not good for complex analysis
Pro tip: Use it to brainstorm ideas directly within your project docs

8. Claude (Anthropic)

Use for: Long-context understanding and analysis
Pros: Can read and analyze very large documents (up to 150K tokens)
Cons: Output is a bit less bold than GPT-4
Pro tip: Use Claude for contract reading, report analysis, and eBooks

9. Leonardo AI

Use for: Game art, concept art, stylized visuals
Pros: Quickly create character designs and scenes
Cons: Niche use – not suitable for traditional photography
Pro tip: Use it to mock up visual concepts for pitches or presentations

10. Descript

Use for: Video editing, podcast production, and voiceovers
Pros: Edit video as if it were text. Automatic transcription and overdub
Cons: Not ideal for advanced video projects
Pro tip: Use it to create snackable video content fast

Step-by-step: How to get started

1. Choose 2–3 tools that fit your needs (e.g. content, video, research)
2. Create an account and go straight to their “Getting Started” section or YouTube tutorials
3. Run a small test case (e.g. write an article, create a video, or brainstorm product ideas)
4. Evaluate the output: Is it fast? Accurate? Useful?
5. Integrate the tool into a specific workflow (e.g. campaign production or sales funnel)

No-code vs coding in AI: now everyone can code

Do you need to learn to code to benefit from AI? No. But will you be able to do more, faster and cheaper, if you can code? Yes.

There are two main paths in the AI world: no-code (drag-and-drop, click-and-run) and code-based (Python, APIs, frameworks).

Here’s what you need to know.

Advantages of no-code AI

Fast access: Get started in minutes without a technical background.
No development: No deployments, GitHub, or error codes.
Drag-and-drop: Build flows using visual builders like Zapier or Make.
Perfect for MVPs: Test ideas quickly without writing a single line of code.

Disadvantages of no-code

Limitations: You’re confined to what the platform allows.
Cost: Many no-code tools quickly exceed $70/month.
Performance: Slower and less flexible for complex tasks.

Advantages of coding (Python, OpenAI API, LangChain)

Full freedom: Build exactly what you need from scratch.
Scalability: Ideal for products, automation, and advanced AI apps.
Access to everything: Use OpenAI, HuggingFace, Pinecone, vector search, and more.

Disadvantages of coding

Time-consuming: You need to learn or hire a developer.
Setup: Requires working with API keys, Python environments, and error handling.
Not for everyone: If you hate coding, don’t do it – it’ll be a struggle.

Concrete examples – when does each make sense?

Use no-code when…

• You want to build a simple lead generator using ChatGPT and Zapier
• You need AI-powered customer support without developers
• You want to validate an idea before investing in development
• You want to automate content flows from ChatGPT → Notion → Mailchimp

Use code when…

• You want to build a custom AI chatbot with memory, vector search, and role logic
• You’re building a SaaS product with AI at its core
• You need full control over functionality and output
• You’re automating large-scale data processing and complex flows

My advice – a pragmatic approach

Start with no-code, learn what works – and go technical when it makes sense.

Many people dive into Python before they even know what they want to build. That’s a waste of time. Flip it: learn the use case first – then the tech.

Examples of no-code tools and low-code frameworks

• Make (formerly Integromat): Advanced AI automations using ChatGPT and other tools
• Zapier: Quick workflows between tools like Typeform, ChatGPT, Slack
• Bubble: Build AI-powered web apps without code – perfect for MVPs
• Replit + LangChain: Low-code framework for advanced AI agents (with some Python)

AI for graphic work and video

If you work in content, marketing, or design – now is the time to pay attention. AI has already revolutionized how we create visuals and video. And it’s moving fast.

You no longer need to be a graphic designer, animator, or video expert to produce content that looks professionally made.

MidJourney – image generation with wow effect

Use for: Illustrated content, product mockups, social visuals, idea development
Strengths: Extremely aesthetic, artistic, and detailed visuals
Weaknesses: No classic web interface – everything runs via Discord
Tip: Use precise prompts with style, lighting, camera, and composition (e.g. “cyberpunk city at night, ultra wide-angle, cinematic light”)

GPT-4 Vision – understanding and generation from images

Use for: Analyzing images, explaining graphs, interpreting UI layouts, brainstorming visuals
Strengths: Understands complex images and explains them in words. Useful for everything from research to UX design
Weaknesses: Not for generating images – but for analyzing and commenting on them
Tip: Upload a screenshot and ask the model to analyze the UI/UX and suggest improvements

Runway – AI video that actually works

Use for: Video production, content creation, greenscreen effects, text-to-video
Strengths: Create full videos from just a prompt – ideal for social media and product demos
Weaknesses: Output is still limited in duration and realism – but improving quickly
Tip: Use “Text to Video Gen-2” to create 4–6 second clips for reels and ads

Sora – the future of AI video

Use for: Advanced video generation (when available)
Strengths: Generates hyper-realistic videos with complex camera movement and storytelling
Weaknesses: Still in closed testing. Not widely available yet
Tip: Follow developments closely – this will reshape the content industry

Other relevant tools

• Canva AI: Generate slides, images, and layouts with one click
• Adobe Firefly: Text and image generation built into Adobe tools – great for professionals
• Pika Labs: A promising alternative to Runway for short, AI-generated video clips

Real-world examples

• An e-commerce site generates product images in MidJourney without photo shoots – saving $4,000/month
• A SaaS company creates onboarding videos with Runway and Synthesia – no editors or filming required
• A consultant builds presentations in Canva AI with autogenerated visuals and design

Step-by-step: create visual AI content in 10 minutes

1. Choose a format: image or video
2. Write a concrete prompt (e.g. “Close-up shot of a smartwatch on a marble table, product photo, high-end look”)
3. Generate using MidJourney or Runway and tweak the output if needed
4. Edit in Canva, CapCut, or Figma for fine-tuning
5. Publish – and test engagement against traditional content

How to learn more about AI

Here’s the honest truth: You won’t get good at AI by scrolling LinkedIn or reading hype posts. You get good by going deep, experimenting, and learning from quality sources.

Top courses (worth the money)

• Complete AI Guide (Udemy): A solid intro to AI, LLMs, prompt design, and real-world applications
• DeepLearning.AI – Prompt Engineering for Developers: Free course by OpenAI + Andrew Ng. Code-focused but hands-on
• AI for Everyone (Coursera): Ideal for leaders who want to understand AI strategically

Blogs and newsletters to follow

• AI Tools by Synthesia: Regular updates on the latest AI tools with examples and use cases
• Ben’s Bites: Daily overview of the most important AI news in a quick-read format
• Superhuman AI Newsletter: Tactical and strategic insights for AI builders and professionals
• Hugging Face Blog: For deeper dives into technical details and what’s under the hood

YouTube channels that actually teach

• Matt Wolfe: Covers the latest tools and trends without hype
• Two Minute Papers: Explains new AI papers at an understandable level
• CodeBullet: More entertainment, but shows what AI can really do

Practical tutorials and playgrounds

• OpenAI Playground: Test your own prompts with full control over model, temperature, and output
• FlowGPT: Community-driven prompt templates for inspiration
• LangChain Docs: For those building AI agents and apps with code

Books to go deeper

• The Coming Wave – Mustafa Suleyman – How AI and tech are transforming the world
• You Look Like a Thing and I Love You – Janelle Shane – A fun and insightful read about AI’s limits and strengths
• Artificial Intelligence: A Guide for Thinking Humans – Melanie Mitchell – A smarter take on what AI is (and isn’t)

Step-by-step: how to learn AI fast

1. Take 1 online course and complete it – e.g. Udemy’s “Complete AI Guide”
2. Subscribe to 1–2 newsletters (e.g. Ben’s Bites and Superhuman)
3. Test 3 AI tools in real use cases
4. Set aside 30 minutes per week for YouTube/experiments
5. After 4–6 weeks: pick one focus area (e.g. visual AI, APIs, or sales automation)

Cases and statistics

It’s easy to talk about AI. But what actually works in practice? Here are real-world examples of companies and individuals implementing AI successfully – with the numbers to back it up.

Real-world cases – how others are using AI

1. Klarna – AI for customer support
   Replaced 700 agents with a single AI chatbot handling ⅔ of inquiries with a 70% resolution rate. Savings: millions annually.
2. National newspaper – AI in journalism
   Uses AI to generate headlines and content fragments for journalists to refine. Result: 30% faster output with no drop in quality.
3. HubSpot – AI in sales
   Integrated GPT into their CRM and saw a 40% increase in email reply rates from AI-generated copy.
4. Solo entrepreneurs and freelancers
   Use ChatGPT, Jasper, and MidJourney to create 3–5x more content without outsourcing. Many save 10–15 hours per week by automating emails, visuals, and videos.

Statistics: Why AI matters

• 83% of companies now prioritize AI strategically (McKinsey, 2024)
• 67% of marketing teams use AI daily – especially for text, SEO, and visuals
• 56% of CEOs expect AI to fundamentally reshape their business models within 3 years
• AI can increase productivity by 40–60% in writing, customer support, and analytics
• Google Trends: Searches for “AI tools” rose over 500% in 12 months

Step-by-step: use cases and data in your AI strategy

1. Find 2–3 cases in your industry – how are others using AI?
2. List manual tasks in your workflow that could be automated
3. Set KPIs: time, output, quality – and test AI for 1–2 tasks
4. Evaluate after 14 days – is it working? Can it scale?
5. Make a short internal pitch – with data – to get buy-in or budget

Conclusion

It’s not about whether to use AI – but how and how fast.

If you’re still on the sidelines waiting for AI to “mature,” you’re already falling behind.

Behind competitors. Behind colleagues. Maybe even behind your customers. I see it every day – not just in marketing, but across all business functions.