Embracing AI: Creating a product content prompt-writing guide for in-house LLM

Problem statement

At Babbel, we have access to Langdock, an LLM, and have been encouraged to use AI as meaningfully as possible to increase efficiency.

How would this look for the Content Design team? Our small team split into different groups to focus on creating artefacts, tools and best practices to engage with AI.

I took on the task to create a prompt-writing  best practices document for creating Babbel product content with LLM.

Method

I referenced various sources such a valuable Button conference talk by Trisha Causley: Content Designer to Prompt Engineer, the UX Content Collective Prompt Checklist, and more.

Outcome

The document I created detailed the following,
as must-dos for product prompt-writing:

  • Role

  • Context

  • Structure & Output

  • Tone

  • Examples

  • Instruction

  • Feedback & Quality Control

  • Final Steps

I included Babbel's Voice & Tone guide, good and bad examples,
types of UI components and information about discovery phase.

This was a project that included a content design colleague, and her next focus was to create a "cheat sheet" with actual prompt copy, templates and sample text.


Next steps

As Babbel increases its AI usage for efficiency, we introduced our work to the entire product design team at a monthly sharing session. Now we are looking at building these prompts into an assistant.