As LLMs conquered the market and early adopters experimented with the tools for writing, the enthusiasm was great: suddenly it was possible to have ChatGPT and Co. formulate supposedly high-quality posts and landing pages with just a few clicks. But since the broad mass of editors and website operators have been using the tool, it's noticeable that the texts are often generic and interchangeable. On LinkedIn, numerous users are already discussing whether the em dash is now a quality feature and how to expose LLM-written texts based on the emojis used.

Yet there is a very simple solution for the typical cookie-cutter ChatGPT results: individual presets. By now, users can easily personalize ChatGPT and provide their own tone of voice as a default setting. For SEO, these settings are worth their weight in gold. Used cleverly, the settings ensure that ChatGPT adheres to best practices for high-quality SEO content.

What does "personalizing ChatGPT" mean in concrete terms?

OpenAI offers an easy way to specifically configure ChatGPT to your own communication style with the individual settings in the user account. The function is part of the "Customize" options and allows you to store your own instructions (prompts) that the LLM automatically considers in every new conversation (at least most of the time).

Possible customizations are:

  • In what tone ChatGPT should communicate (e.g., factual, creative, direct)
  • What professional depth or target audience is expected
  • Whether certain terms should be used or avoided
  • How the response structure or formatting should look (e.g., lists, short paragraphs, markdown)
  • Information about the user themselves
  • Memory about the person

These pieces of information thus act like a kind of "invisible prompt" that is automatically considered in every new conversation.

How do you store personal instructions?

Under the settings in the ChatGPT interface (bottom left) you'll find the "Personalization" section. There you'll find the setting options mentioned above:

What tone should ChatGPT use?

In the dropdown, the tool offers numerous options to define the language style. The options range from professional or friendly style to a nerdy or cynical writing style. The style can be further refined at a later point.

What individual instructions should the LLM observe?

In the free text field, additional specifications can be stored that go beyond the tone. These can be small rules, personal preferences, or quality standards that the model should automatically keep in mind. These instructions work like a small compass: They ensure that responses sound less like standard ChatGPT and are closer to what is really needed in everyday life.

Typical points for personalizing ChatGPT here:

  • How responses should look: rather short, rather detailed, with examples, in list form, or clearly structured.
  • What's important: fact-based content, honest hints in case of uncertainties, no made-up stuff.
  • Linguistic subtleties: prefer or avoid certain terms, typical spellings, desired formats.
  • Everyday context: industry, role, or recurring topics that are relevant to the work.
  • Small do's & don'ts: no unnecessarily formal style, no filler sentences, no repetitions.

Tip: Tired of the typical ChatGPT blah-blah? In our article on typical ChatGPT phrases, we've collected a lot of inspiration for fixed rules for the individual instructions.

"About you"

In this section, users can disclose "private" information about themselves. In the "Nickname" field, there's the option to store your own name. ChatGPT then addresses the user by this name in chats. Thus, the user can also personalize the name in ChatGPT.

Your own profession can also influence how ChatGPT generates responses in the future. A social media manager most likely uses a different writing style and different word choice than an auditor. A kindergarten teacher needs different ideas for games than a vocational school teacher. That's why it makes sense to store more detailed information about the person who actually uses the tool.

What goes in the "More about you" field?

The "More about you" field is intended to give the model a bit of background knowledge about the motivation and goals, so that responses better fit your own context. So this isn't about private matters, but about information that helps the LLM respond more relevantly. Typical content is, for example:

  • Work environment: industry, role, typical tasks.
  • Goals: e.g., creating high-quality content, working more efficiently, preferring certain formats.
  • Recurring topics: such as SEO, web development, content, marketing, project work.
  • Helpful framework information: preferred tools, target audiences, writing style preferences.
  • Organizational matters: e.g., whether short summaries are often needed, whether markdown is helpful, or whether texts should be directly usable.

What are memories?

Memories are probably among the most underestimated presets. At the same time, they're a bit tricky because they can't be set quite obviously. But with a small hack, it's still possible: We don't have to ask ChatGPT to store memories about us. The tool does that all by itself.

With recurring information, ChatGPT gradually stores information about us and our work. Users can check which ones under "Manage." Don't be startled: A list opens with all sorts of notes about us that have accumulated over time. It's worth going through the list once and deleting notes that aren't always relevant or might be outdated by now.

What many don't know: If you prompt ChatGPT in a chat to save a specific memory, the LLM does that too. It takes a small moment until a notification appears that the memory was successfully saved. An example could be: "Save the memory that I am an SEO manager and specialize in content SEO."

At some point, however, the tool has had enough too. ChatGPT indicates when the list is (almost) full. From that moment on, it no longer saves any further memories. That's why only truly relevant information should be stored here.

CustomGPT for individual applications

In addition to the personal basic settings, there is now another way to tailor ChatGPT to very specific tasks: CustomGPTs. They function like small, custom-made versions of ChatGPT – ideal for recurring tasks, firmly defined workflows, or internal processes. Instead of having to explain each time what is needed, you can store your own rules, data, and prompts directly in the GPT.

For a detailed explanation of how CustomGPTs are structured, when they make sense, and how companies can work with them, we've published a separate article. It goes much deeper into detail there – including advantages, use case scenarios, and practical examples.

Click here for the article "Creating CustomGPT".

Conclusion: Small setting, big impact

Personalizing ChatGPT raises the quality of your own texts to yet another level. Anyone who regularly works with AI should use these functions purposefully to save time, remain consistent, and achieve better results. Because generative AI is only a real added value when it understands your own style.

Tired of mediocre content? We know the hacks for SEO-optimized and target-audience-appropriate content.