Skip to main content Skip to page footer

Prompting Best Practices

In this article, we will show you what makes good prompting, which best practices have proven effective – and how the Prompt Manager of the AI Suite helps you manage your prompts efficiently and in a structured way.

Why Good Prompting is Crucial

A KI model like ChatGPT, Claude, or DALL·E works with what you provide it. Poor or unclear prompts often lead to irrelevant texts, inappropriate tones, or results that simply miss the mark. Good prompts, on the other hand, help to structure content clearly, consider style guidelines, pinpoint complex requirements, and save a lot of time in everyday tasks.

Prompting Best Practices – What You Should Consider

1. Formulate Precisely

The more specific a prompt is, the better the result will be.

Poor: "Write a text about TYPO3"

Better: "Write a 3-4 line introductory text about the advantages of TYPO3 for medium-sized businesses"

2. Assign a Role

A clear context leads to more targeted responses. Example:

"You are an SEO expert with 15 years of experience creating content for a digital agency."

3. Structure Defines the Output

Specify in what format you want the answer:

"Create a table with three columns: Feature, Advantage, Target Group"

4. Provide Style Guidelines

Exemplary criteria:

  • Tone: professional, casual, promotional
  • Language level: simple, technical, academic
  • Length: "max. 300 characters"

5. Provide Context

If prompts are to build on existing texts or data, provide that context:

"Use the following paragraph as a basis and write a meta description under 160 characters."

6. Break Down Complex Prompts

A common mistake is packing too many requirements into one prompt. Better is:

  • Prompt 1: Generate content
  • Prompt 2: Check tone
  • Prompt 3: Format for SEO

7. Test Variants

Even small differences in the prompt text can have a big impact.

Avoid These Mistakes

  • Vague or ambiguous formulations
  • No defined role or target group
  • Missing output format (list, table, HTML)
  • Too many requirements mixed in one prompt

Prompt Engineering with 'Few-shot' Examples

Results can be significantly improved through example prompts (Few-shot Learning):

"Example 1: ...

Example 2: ...

Now write a similar text about ..."

This is particularly effective for tasks such as:

  • Product descriptions
  • FAQs
  • Social media posts

The Prompt Manager of the AI Suite

With the Prompt Manager in the AI Suite, prompts can be managed centrally, filtered by scopes, and saved as templates. This not only saves time but also creates reusability – for individuals as well as entire teams.

Especially helpful: Placeholder support. This creates dynamic templates:

Create an introductory text for the content element [Element_Name] in the style of [Style_Guideline] with a maximum of [Character_Count] characters.

These templates can be easily duplicated and adapted to different use cases.

Conclusion: Invest in Your Prompts

A good prompt is like a good briefing – the clearer, the better the result. With the AI Suite for TYPO3 and the Prompt Manager, you have a tool that helps you not only write prompts but also use them strategically and reusable.

This way, you not only get better content – but also more consistent workflows, scalable templates, and maximum efficiency.

Our Tip: Create your own prompt templates in the Prompt Manager and continuously optimize them based on your use cases.