SOMA Systems
Intelligence layer — within SOMA Systems

SOMA Engine

The adaptive intelligence behind the interface.

It decides whether, and how little, to act — and whether no action is the better response.

SOMA Engine is the adaptive intelligence layer within SOMA Systems. It supports the product architecture across sensing, interpretation, anticipation, intervention, and adaptation — and it carries the governance that keeps each response proportionate.

A soft golden field of flowing particles and threads of light against warm dark
Fig. 01 — The layer behind the signal

Intelligence as restraint — the work is mostly deciding when not to act.

What it is

The judgment the interface defers to.

An interface like Somatag senses and signals; it does not decide. SOMA Engine is where the deciding happens — comparing a signal against a personal baseline, weighing context and confidence, and choosing the least-burdensome response that might help. Often, that response is to stay quiet.

What it weighs, every time
Personal baseline Context Inference confidence Intervention history User preference Safety policy Whether no action is better
The adaptive loop

Five movements, and a way back to the start.

01

Sensing

Receive selected physiological and contextual signals, with their uncertainty intact.

02

Interpretation

Estimate meaningful change against a personal baseline — not a universal threshold.

03

Anticipation

Weigh confidence and context to find a window where support might still help.

04

Intervention

Choose the least intrusive response that may be useful — or choose silence.

05

Adaptation

Observe whether it helped, burdened, or did nothing — and learn from recovery.

Recovery is part of the loop. An intervention that helped and one that intruded are different lessons.

How it governs itself

Confidence before action

Low confidence reduces system behavior — it does not manufacture certainty.

Minimum necessary response

The least intrusive intervention likely to be useful, and never more.

No action is valid

Adaptive does not mean constantly active. Silence is a first-class decision.

Support without diagnosis

Observable physiology is never translated into unsupported emotional labels.

User control

The wearer can understand, pause, reverse, or disable system behavior.

Personal baselines

Meaningful change is individual and temporal — read against the person, over time.

The entity architecture

Where the Engine sits.

Christine Galligan Founder & principal
Practice · field labKutuhala StudioResearch, foresight, and design — the active practice. Nonprofit · evidenceUmwelten InstituteBuilds the evidence. Does not commercialize. Commercial · systemsSOMA SystemsBuilds and commercializes the systems.

SOMA Engine is a SOMA Systems asset. It should not be described as a nonprofit asset unless counsel explicitly structures it that way.