FrameworkWhy we built our own prompting framework
A seven-step framework for reproducible AI results in a business context.
Developed in-house by Radical Innovators. Freely available, including a cheat sheet as PDF.
Starting point
In companies, AI adoption rarely fails on the technology — it fails on operation. Prompting stays trial and error: what works for one person cannot be passed on, repeated, or built into a process.
Approach
Seven named steps — R·A·D·I·C·A·L — that turn a vague request into a reproducible working instruction: Role, Ask, Data & context, Industry, Constraints, Answer format, Learning loop. The decisive step is industry context: a prompt without it returns generic answers.
How it came about
The framework grew out of client project practice. Workshops kept revealing the same pattern: individual employees achieved impressive results with AI — but could not explain why. Nothing about that scales, and nothing transfers into a process.
Rather than assembling yet another collection of prompt templates, we named the structure underneath. Seven steps, each with a concrete action, in an order you can remember. The result is deliberately not a product but a method: freely available, explained at length as an article, and printable as a one-page cheat sheet.
We publish it openly because it makes our work better, not because it replaces it. A team that can prompt in a structured way asks better questions in projects.
Scope
- —Method development from project practice
- —Framework systematics (7 steps)
- —Editorial development in German & English
- —Cheat sheet as print-ready PDF
Technologies
Frequently Asked Questions
What does RADICAL stand for?
For the seven steps of the framework: Role, Ask, Data & context, Industry, Constraints, Answer format, and Learning loop. Each letter stands for a concrete action, not a buzzword.
Does it cost anything to use?
No. The framework is freely available — both the in-depth article and the cheat sheet as a PDF in German and English.