Prompt Quality.
Demonstrator for structured prompts. Pick one of five curated examples and see how structural maturity expresses itself across the five CRAFT dimensions: Context, Role, Action, Format, Tweaks.
Open the tool→Five tags that signal a good prompt.
A simple mnemonic: CRAFT. Each letter stands for one dimension you either set deliberately in your prompt — or deliberately leave out.
Context
Which industry, which regulatory frame, which constraints apply? Context decides whether the AI answers with the right understanding. Example: "Regulated banking, BaFin-relevant, EU AI Act high-risk class." Important when domain knowledge drives answer quality.
Role
Who is the AI — and to whom is it speaking? A clear role saves clarification loops. Example: "You are an experienced test manager in banking." Important when the answer needs depth that drifts to generic without a role.
Action
What exactly should the answer contain? A concrete task beats ten vague wishes. Example: "Build a prioritized top-5 risk list." Always important — without a clear action the AI heads somewhere you did not intend.
Format
What shape should the answer take? Markdown table, bullet list, prose, JSON, code block? Format upfront saves reformatting later. Important when the answer feeds another system or has to fit a specific layout.
Tweaks
What constraints, what tonal rules, what fallback behavior? Max length, allowed phrases, source citation, ask-back vs assume. Important when the AI would otherwise default to generic, too-long, or too-confident answers.
Try it out.
The rating runs entirely in your browser. Your input does not leave the window.
What you notice as you click through the examples.
How we use Prompt Quality inside a team.
We use the CRAFT framework in our AI workshops as a shared language for prompt quality. The tool here serves as a demonstrator. For a deeper dive there is the AI fundamentals workshop, the AI-as-QA-tool workshop, or the methodical AI-evaluation workshop for test-leading roles.
Workshop: AI fundamentals
Four-hour workshop for everyone working with AI in a structured way for the first time. You get to know the most important models and tools, practise prompts and role briefings, and leave with a clear view on use cases and risks in your daily work.
→
Workshop: evaluating AI
Two-day workshop with a concrete procedure model for test leads who have to evaluate AI-assisted features or full AI systems. Together we set up a local AI environment, build a small demo application with an integrated AI assistant using Claude Code, and run through the test categories functional, guardrails, adversarial and localisation on it. Prerequisite: a solid understanding of AI, development environments and test methodology (more on this in the FAQ).
→Prompts that hold up in everyday work.
If you want to support your team in writing prompts in a structured way, talk to us. Workshop, one-to-one coaching or simply a first conversation.
Book a call→