ORBIS Prompting Guide for Makers
Sectors: Energy suppliers, Non sector-specific, Manufacturing industry, Construction supplier industry, Steel / Steel Service Center / Metal industry, Electro and electronic industry, Mechanical and plant engineering, Consumer goods / retail, Automotive supplier industry
Processes: Project Management, Change Management, IT, Service & Support, Sales, Marketing, Data Management
ORBIS Prompting Guide: How to write perfect prompts with the right AI instructions
The ORBIS Prompting Guide for Makers shows you step by step how to structure your prompts so that generative AI models understand your requirements, from initial idea through to finished solution. Learn how to ask the right questions, structure tasks effectively, incorporate feedback loops and sustainably increase the quality of your deliverables.
Understanding generative AI
The possibilities offered by generative artificial intelligence range from automated text creation and application development through to intelligent data analysis. However, to use this technology effectively, you need to know the basics.
What exactly is generative AI? How do AI models like Microsoft Copilot work? And why are precise prompts essential for good results?
Introduction to generative AI
Generative AI is a form of artificial intelligence that can create new content such as text, images, code and analyses based on existing data and speech input. It understands context and processes speech input (prompts) to independently generate answers, suggestions and even whole documents. Well-known examples of this are Microsoft Copilot, ChatGPT and DALL-E.
In a business context, generative AI is being increasingly integrated into Microsoft applications such as Power Apps, Power Automate, Power BI and Dynamics 365 to automate processes, accelerate decisions and create hands-on content.
Generative AI-Basics
The functionality of generative AI is based on what are known as large language models (LLMs). These models have been trained on vast amounts of text data and recognize patterns, connections and meanings.
Instead of ‘looking things up’, they generate content based on probabilities, in other words, what makes the most sense in a certain context. The important thing to know here is that the output can only be as good as the prompt. A clearly formulated prompt with context, objective and format specification is key to how precise and helpful the AI's response will be. So, if you understand the basics and have an effective command of prompting, you can use generative AI in your everyday work in a focused and productive way.
The challenges of generative AI
Generative AI like Microsoft Copilot and other large language models are changing the way we work, write, analyze and develop. These new opportunities also bring new challenges.
The quality of the results largely depends on the quality of the prompt entered. A lack of structure, unclear tasks, or questions that are too general quickly lead to imprecise, incorrect and unusable answers. Plus, many users still lack experience in dealing with generative AI. Proper application, an understanding of limits and potential, and efficient use in day-to-day work are all things that are often not yet embedded.
This is where the ORBIS Prompting Guide comes in! It helps to master such challenges with a structured, clear approach and best practices for better results
Understanding prompts and prompt engineering
Systems with generative artificial intelligence (AI) are designed to deliver certain results based on the quality of the queries (prompts) they receive. So that these systems can understand and answer their simple or more complex queries, certain factors need to be taken into account. This is where prompt engineering comes into play.
What is a prompt?
To understand prompt engineering, you first need to know what a prompt actually is. A prompt is a specific instruction or question that a user enters to guide an artificial intelligence (AI) such as Microsoft Copilot. Prompts give the AI clear information and content about what it is to do as well as the context and format in which the result is expected. The more precise and structured a prompt's formulation is, the better and more relevant the AI's responses will be.
What is prompt engineering?
Prompt engineering refers to the targeted creation and optimization of prompts to get the most precise, helpful and high-quality results possible from an AI. This isn't a matter of asking questions, but of clearly structuring tasks, providing relevant information and continuously refining results through feedback loops. The aim is for the AI to implement complex requirements efficiently, in an understandable way and with pinpoint accuracy.
How does prompt engineering work?
Using the right strategy not only increases the quality of AI results but also saves valuable time in development and work processes. Learn what it takes to write good prompts and how you can achieve better results with just a few simple methods.
How do you write a good prompt?
Good prompts are clear, precise and structured. A good prompt contains the following elements:
- Clear task definition: What exactly is the AI supposed to do?
- Specific details: Time frame, target group, region, preferred format, etc.
- Positive or neutral wording: Avoid negatively worded questions to promote balanced results.
- Context and references: Refer to templates, examples and sources where necessary.
- Step-by-step logic: Break down complex tasks into smaller sub-steps.
- Incorporate feedback loops: Allow pauses for queries and corrections.
What are the advantages of prompt engineering?
Targeted prompt engineering delivers significant advantages:
- Greater quality and relevance of AI results
- More efficient work processes through precise control
- Consistent results, including with complex tasks
- Increased control over the structure and style of responses
- Time savings thanks to clearly formulated instructions and less reworking
Well-formulated prompts turn your AI from a tool into a real sparring partner.
Prompting for Makers – what can you expect?
Theory is great but real-world application is better. We provide you with specific examples so that you know straight away what you need to consider when writing a good prompt. You will learn which typical mistakes to avoid and how you can significantly improve the quality of AI responses by using clear wording, specific details and a sound structure.
Because a good prompt is more than just a question. It's a clearly formulated task with a purpose, context and format specifications.
You will learn:
- How to formulate clear and specific prompts
- Why positive and neutral questions lead to better results
- How to integrate reference documents and feedback loops into your prompts
- What role persona definitions, intermediate steps and iterations play
- Which practical use cases you can cover with the right prompts
No matter whether Power Apps, Power Automate, Power BI or Copilot Studio, the real added value of AI agents reveals itself when you formulate your prompts the right way.
