Home / Glossary / CDT — Context Design Thinking

CDT — Context Design Thinking

A four-phase methodology for designing context in AI systems — Audit, Architect, Construct, Integrate. The prompt is the precise instruction; the context is the environment that defines how it is interpreted.

One of the three disciplines of AI-Human Thinking. A four-phase methodology for designing context in AI systems: Audit (inventory and relevance classification of available information), Architect (design the hierarchy — critical information at the beginning/end, support in the middle), Construct (build the system prompt with structural patterns), Integrate (integrate with the system in production).

The central distinction: the prompt is the precise instruction; the context is the environment that defines how the instruction is interpreted.

Foundational analogy: Context is the computer screen — the complete environment where everything happens. The prompt is the mouse movement and what is typed — the precise instruction within that environment.

Context & Strategy

How it was born

Developed in parallel with PDT, from the need to conceptually separate two types of work in AI: what is in the prompt (precise instruction) and what is in the context (everything that informs and frames). Most existing methodologies treat the two as synonyms — CDT formalises the distinction.

What it does in practice

Used in the design of AI agent systems, in the construction of system prompts for SaaS products, and in teaching AI-Human Thinking. The prompt/context separation is one of the most citable conceptual contributions of the Bitsapiens ecosystem.