AI Prompting – The Ultimate Starter’s Guide for ChatGPT and Midjourney

Tobias Zehnder

Co-Founder / Partner

Powerful, useful, tricky: Generative AI in everyday marketing and communications holds a lot of potential and seemingly endless possibilities. The key: prompting – a word that has been around for a long time, but has only recently entered our everyday vocabulary. Social media is full of prompting tips, and the selection is huge. We’ve compiled the two most important basic rules for good AI prompts and the ultimate starter’s guide for ChatGPT and Midjourney.


The two most important basic rules for good AI prompts – whether for images or text – are simple but essential:

  • Give clear and concise instructions
  • Give context so that the tool understands the task

After that, it’s a matter of iteratively improving the prompts to get to the desired result.

Simple image of boy sitting in a treehouse


Simple prompt: 

kid in a tree house

Complex image of boy sitting in a treehouse

Complex prompt:

[kid]::1 in a gigantic, playful [tree house]::2, vibrant colors, dramatic lighting, photo-realistic, landscape shoot, daylight, mountain scenery --ar 1:1

Interested in a deep dive? The following two prompting guides will give you an ideal springboard for working with ChatGPT and Midjourney.


  • Tone: Specify the desired tone (e.g., formal, casual, informative, persuasive).
  • Format: Define the format or structure (e.g., essay, bullet points, outline, dialogue).
  • Actas: Indicate a role or perspective to adopt (e.g., expert, critic, enthusiast).
  • Objective: State the goal or purpose of the response (e.g., inform, persuade, entertain).
  • Context: Provide background information, data, context for accurate content generation.
  • Keywords: List important keywords or phrases to be included.
  • Limitations: Specify constraints (or minimum requirements), such as word or character count.
  • Examples: Provide examples/references of desired style, structure, or content.
  • Audience: Specify the target audience for tailored content.
  • Language: Indicate the language for the response, if different from the prompt.
  • Citations: Request inclusion of citations or sources to support information.
  • Pointsofview: Ask the Al to consider multiple perspectives or opinions.
  • Counterarguments: Request addressing potential counterarguments.
  • Terminology: Specify industry-specific or technical terms to use or avoid.
  • Analogies: Ask the Al to use analogies or examples to clarify concepts.
  • Quotes: Request inclusion of relevant quotes or statements from experts.
  • Statistics: Encourage the use of statistics or data to support claims.
  • Visualelements: Inquire about including charts, graphs, or images.
  • Calltoaction: Request a clear call to action or next steps.


Important for Midjourney: Always start prompts with "/imagine".

  • General Style: Street style, editorial style, food photography, 1990s punk style
  • Composition: Off-center closeup, medium-full side angle, full body shot, two-shot
  • Medium: Photo, film still, illustration, sketch, sculpture
  • Film Type/Camera: Kodak Gold, Agfa Vista, Canon EOS 5D, Kodachrome
  • Subject Description: A woman walking, a man talking, a dog running, a daisy flower
  • Subject Styling: Fashion descriptions, subject details, anything that details the appearance
  • Environment: New York, Central Park, Ancient Egypt, Underwater
  • Lighting: Natural lighting, dramatic lighting, studio lighting, off-camera flash, overcast
  • Atmosphere: Misty, smokey, steamy, foggy
  • Mood: Dreamy, elegant, spooky, romantic, luxurious

Sources: Manuel Nappo (Hochschule für Wirtschaft Zürich), Nick St. Pierre @nickfloats (Twitter), ChatGPT