Prompt engineering: the term, the job profile and its use in SEO and SEA

Prompt engineering is the art of precisely wording instructions, known as prompts, for AI models such as ChatGPT. It requires the ability to write clear and specific instructions in order to obtain the desired answers or solutions. This process calls for an understanding of what the AI can do, so that it can be steered in the right direction. Prompt engineering is essential for getting targeted and relevant answers from AI models and is used in various applications such as text generation, data analysis and problem-solving. It therefore also demands creativity, experimentation and fine-tuning to get the desired results out of the AI.

A prompt engineer is, as you may have guessed by now, responsible for writing the instructions for the AI. This means the prompt engineer is presented with a problem and has to solve it with the help of the AI, or write a prompt template that others can use to have a particular output produced by an AI. There are therefore many possible uses for prompt engineering.

AI can, for example, make programming easier, because it can analyse the code entered and find errors or help to write new code. In business, too, AI is now used to simplify data analysis. AI is also used in healthcare customer service, for instance. Last but not least, online marketing can benefit from prompt engineering, for example in SEO and SEA.

Because prompt engineering is quite demanding, it is being hailed as a new career path. In the United States in particular, prompt engineers were offered high salaries for a time. The AI company Anthropic in San Francisco, for example, advertised a prompt engineer position in 2023 with a salary of up to 335,000 dollars (around 310,000 euros at the time). To earn that much, however, candidates need to demonstrate sufficient experience and expertise.

In Germany, the job title prompt engineer has not yet become established, but courses from private providers are already offered online. There, anyone interested can learn how to work with AI. This is certainly useful as an additional qualification, especially in the industries mentioned above. There is no fixed set of requirements for the job of prompt engineer. Candidates therefore have to show initiative and familiarise themselves with the subject matter of the company or industry in question. So prompt engineering can only be learnt in part. Companies place more value on other qualifications, such as experience in programming or machine learning (the foundation of AI).

Search engine optimisation is about good content that is tailored to users and to search engines. AI can help here, for example by suggesting the structure for a blog post or a service page that the copywriter can work from. Having posts written entirely by an AI, on the other hand, tends not to pay off. Google rates the usefulness of content and not the tool used to create it, but it can treat pages produced at scale without added value as spam. Users, too, notice when content is not authentic. Copywriters therefore need to be just as good at writing prompts as they are at assessing the output critically.

Prompt engineering can be particularly useful for product descriptions. To do this, prompt engineers have to both write a perfect prompt and prepare the relevant database from which the AI draws its information about the products. AI can likewise be used to produce smaller items such as the meta description or accompanying social media posts for finished blog posts. This works because the AI can draw on content written by people and combine it with the SEO strategies it has learnt.

AI tools such as ChatGPT can also be used in search engine advertising. Through targeted interactions, advertisers can create tailored campaigns. Precise requests have to be worded using prompt engineering, for example to identify suitable keywords, develop creative ad variants or analyse the effectiveness of ads. Competitor analysis and performance analysis are also possible using prompt engineering and AI, and could streamline some processes in online marketing agencies.