TALAMANA · THE AI LITERACY MAP FOR ARCHITECTURE AND DESIGN · The Brief · AGE 18—20 · POSITIONAL · METHOD
The six-layer reasoning brief
A reasoning brief has six layers, and the model fills any layer you skip.
When to use
Any time you hand a reasoning model real design work — not a quick question, but a task whose answer you must stand behind.
The method
Six layers, in order: intent (what the work is for) · context (what the model must know) · constraints (what must not move) · criteria (how you will judge the result) · output form (what shape the answer takes) · iteration loop (what happens after the first pass). Write all six. Skip a layer and the model fills it with an assumption — and its assumption, not yours, then shapes the work.
Watch for this
Writing constraints but no criteria. The model then knows what it cannot do, but not what "good" means — so it optimises for plausible instead.
The Lab's note
Six layers is our method, refined through teaching. It is not the only workable scaffold. It is the one we can defend layer by layer.
Try it
Take a brief you actually sent this week. Mark which of the six layers it contained. Rewrite it with all six, run both, and put the two results side by side.
Prove it
Show one brief where each layer is present and you can point to where each one changed the outcome.
How it works
A reasoning model plans before it answers. What it plans WITH is whatever your brief supplies. The six layers map to the decisions the model must take anyway; the brief simply decides who takes them — you or it. This is the text-engine twin of the seven-slot visual brief elsewhere on this map.
What this idea builds on
What this idea opens up
- Nothing yet names this as a foundation.
Sources
Open this idea on the map · The complete map · Logika · RBDS AI Lab, India · revised every edition.