Recipient Intelligence
Recipient Intelligence is the layer that gives an AI a durable model of the recipient — and adapts every message to them. It's the EQ layer for AI.
What it is
Recipient Intelligence is a capability layer that sits between an AI agent and the send action. Before a message goes out, it answers three questions the model can't: who is this for, how do they prefer to be communicated with, and what does the relationship need right now. It is the direct answer to Recipient Blindness — the structural gap where a system writes fluently but has no model of the person reading.
"EQ layer" and "emotional intelligence" describe the effect. Recipient Intelligence is the precise name for the mechanism: a model of the recipient, applied to the message.
The same intent, two recipients
Recipient Intelligence isn't tone. It changes structure — order, length, what leads, what's cut. Here's one intent written for two different recipients:
Same goal, same model, different structure — because the recipient is different. A system with no model of the recipient can't make this choice, so it sends the average to everyone.
How it works
Recipient Intelligence rests on three primitives:
Underneath, recipients are mapped to a taxonomy of communication archetypes — Dragon, Oracle, Sphinx, and nine others — each with its own adaptation guidance. The archetype is what makes adaptation repeatable instead of a one-off guess.
How it's measured
A category needs a standard. The Recipient Intelligence Index (RII) scores systems 0–10 across five dimensions — recipient state persistence, adaptation depth, feedback integration, archetype differentiation, and relationship-health tracking — to separate genuine Recipient Intelligence from the appearance of it.
Build with it
Recipient Intelligence is an API call. Profile a recipient, then adapt a draft to them:
/context/person and /adapt/message in the docs.