The entire article is also available as a podcast episode:
Attention! The podcast was created entirely by my AI-Assistant based on my contribution – no guarantee for incorrect content.
The use of artificial intelligence (AI) has developed rapidly in recent years. Generative AI models in particular, such as ChatGPT, have opened up new possibilities for processing information, developing ideas and automating tasks. However, in order to exploit the full potential of these models, a targeted strategy for prompting – i.e. formulating inputs – is crucial. In this article, I will introduce you to three key prompt methods and use specific examples to show you when and how you can best use them.
Don’t worry – many of the previously known prompt methods will be retained. However, new methods are being added and some of the old methods can be shortened.
The three central prompt methods
1. conversational prompting
Conversational prompting is particularly suitable for classic GPT models and reasoning models. You use the AI as an interactive sparring partner with whom you develop ideas, work out new concepts or discuss complex issues.
When does this method make sense?
- If you have a new task ahead of you and are unsure how to start.
- If you want to brainstorm with the AI.
- If you have a question for which you do not yet have sufficient background knowledge.
Example prompt:
“I want to develop a marketing strategy for a sustainable fashion company. What aspects should I take into account? Ask me questions to refine my ideas.”
Here, the AI not only responds with an answer, but also asks you specific questions to further develop your concept.
2. assistant prompting
With assistant prompting, you use the AI as a helper for clearly defined tasks. This method works particularly well with classic GPT models and helps to delegate repetitive or time-consuming tasks. To achieve the best results, you should follow a few basic prompt rules:
- Assign a role: Start the prompt with a clear role description, e.g. “You are an experienced copywriter for marketing campaigns.”
- Describe the task in detail: Give as many relevant details as possible so that the AI understands exactly what you expect.
- Set expected result: Defines the desired format or structure of the output, e.g. “Write an email with a maximum of 200 words.”
- Provide sample data or contexts: If possible, provide sample texts or specific contexts to improve the quality of the response.
- Give additional instructions: If necessary, specify certain requirements, e.g. “Use a friendly but professional tone.”
An example prompt that implements these rules could be:
“You are an experienced business copywriter. Write a professional e-mail to a customer who has enquired about a consulting offer. The customer is interested in a long-term collaboration. The e-mail should be polite, clear and professionally worded. Be sure to highlight the key benefits of our offering and highlight the next steps.”
You can also find more prompt tips in my prompt guide.
By following these prompting rules, you can significantly improve the effectiveness of your requests.
When does this method make sense?
- If you already know exactly what needs to be done but want to save time.
- If you want to automate a task that you would otherwise do yourself.
- When it comes to generating text formats, translations or summaries.
3. reasoning-prompting
Reasoning prompting is mainly used for highly complex tasks and is only suitable for advanced reasoning models. This method uses the advanced logical capabilities of AI to support well-founded decisions or analyze complex problems. To realize the full potential, you should use structured prompts that include a clear task, defined expectations and logical chains of reasoning.
- Formulate a detailed problem definition: Give the AI a clearly defined task with relevant background information.
- Ask targeted questions: Instead of open-ended questions, formulate specific requirements, e.g. “Create an argument with three pro and three con points”.
- Define expected output format: Specifications such as “Create a systematic decision overview with tables” help the AI to deliver results more precisely.
Example prompt: “Analyze the economic impact of rising inflation on the real estate market. Consider economic, social and political factors. Present your analysis in a structured overview with clear conclusions and options for action.”
You can find even more tips on reasoning prompting in one of my last posts on the use of Deepseek and Co.
When does this method make sense?
- If you need data-based analysis or decision support.
- If you want to think through complex relationships systematically.
- When it comes to in-depth conclusions that go beyond simple pattern recognition.
Example prompt:
“Analyze the following market trends for the automotive industry and derive possible future developments for electric vehicles. Take geopolitical, economic and technological factors into account.”
Here, AI uses its reasoning-based knowledge to carry out in-depth analyses.
Comparison of prompt methods
The following table provides an overview of which methods are suitable for which use cases.
| Method | Field of application | Advantage | Example prompt |
|---|---|---|---|
| Conversational prompting | GPT models & reasoning models | Ideal for brainstorming and sparring | “I want to develop a new business idea. Ask myself specific questions to refine my idea.” |
| Assistant prompting | Classic GPT models | Efficient delegation of clearly defined tasks | “You are an experienced copywriter. Write a professional email to a customer who has inquired about a consulting offer.” |
| Reasoning-Prompting | Only for reasoning models | Solving complex problems with a logical approach | “Analyze the economic impact of rising inflation on the real estate market and create a structured decision overview.” |
Conclusion: Finding the right prompt method
Choosing the right prompting method depends on your goal and the model in question. Conversational prompting is ideal for open discussions and brainstorming, assistant prompting enables efficient work, and reasoning prompting brings real added value to complex issues. Try out the different methods and adapt your strategy to your needs – this is how you get the best out of AI-supported tools!
Please note that reasoning models currently still consume a lot of energy and time and are therefore usually very cost-intensive. So if you don’t need the advantages of this method, choose a different model and a different prompt method.
Further questions
Has this article given you food for thought and do you have any further questions? Or are you looking for general support with the use of AI, ChatGPT, Deepseek and chatbots?
I am always happy to receive your messages, preferably by WhatsApp message or e-mail.
This article is also available as a podcast episode
Attention! The podcast was created entirely by my AI-Assistant based on my contribution – no guarantee for incorrect content.
