The co-thinking guide
“Human-AI Co-Thinking in Action” is a practical guide to working with AI through dialogue. It requires no technical background.
An emerging form of literacy
For decades, controlling powerful technology meant learning to code. Human-AI co-thinking works through communication: systematically shaping how AI behaves, compensating for its limitations, using its strengths and keeping your own judgement engaged.
The guide shows how to create AI-processes: specialised assistants you design through conversation for a specific type of work. Each one lives in files you keep, read and refine over time. For example a brainstorming partner that challenges your assumptions, a writing coach that knows your style or a project planner that tracks your decisions.
All it takes is clear communication and a willingness to engage with a new kind of interlocutor.
Read or listen to the guide
The guide dates from October 2025.
- Human-AI Co-Thinking in ActionPDF, 63 pages, 41 MB
- The text of the guideMarkdown, about 12,600 words, on Google Drive
- The podcastOn Podbean
Written by Charles-Edouard Bardyn, Vice-President and Chief Scientific Officer of AI Swiss.
What the guide covers
After an introduction to AI and co-thinking, the guide proceeds in nine steps:
- Create your first AI-process.
- Use it.
- Understand its components: instructions, context, memory and scripts.
- Refine it over time.
- Build a library of AI-processes.
- Create an orchestrator for that library.
- Work through the orchestrator.
- Manage the library: upkeep, versions, security and data protection.
- Scale to an organisation with a knowledge hub.
Frequently asked questions and a look at where co-thinking fits among AI tools complete the guide.
For a team or a school
BASE applies this method to a team: its way of working lives in files it can read, correct and pass on. For schools, the AI Swiss white paper describes the approach.