A clinical context layer is a persistent, continuously maintained model of how a patient's facts relate to one another: records, observations, device signals and clinician concerns. Applications draw the context they need from it instead of rebuilding it on every request. Clinical Context AI (IMI) is specifying one. The system proposes; a clinician decides.
Status
Specified · Prototype next · Not in clinical use
Clinical Context AI (IMI) proposes; a clinician decides.
Healthcare keeps information in two places. Systems of observation, such as imaging, labs and monitoring, answer one question and go quiet. The electronic health record keeps what was decided. Neither was built to hold how the facts relate to one another.
That gap matters more as observation becomes continuous: patient state is becoming observable faster than it is becoming coherently interpretable. A clinical context layer is the model that holds the relationships between facts and keeps them current.
Not an electronic health record.
It draws on the record and can feed insight back. It does not replace it.
Not a diagnostic or decision-making system.
It never asserts a clinical fact on its own, and no clinical action is automated.
Not a chatbot or an AI assistant.
It supplies maintained context to applications that reason over it.
Not a data warehouse.
It holds relationships and their changing significance, not stored data for later analysis.
Not in clinical use.
It is specified, with a prototype next.
As described in the whitepaper. This is the specified design, not a deployed system.
An event arrives
A lab result, a medication change, a device reading or a note enters from a device, an application or a system of record.
It is read against what is known
The existing model gives the basis for interpreting it, alongside what clinicians have said matters for this patient.
The model is updated
New relationships are added, earlier interpretations revised, and uncertainty kept as uncertainty.
Significance is evaluated
Conditions that clinicians and domain experts defined are checked. If one is met, the layer can raise it for attention.
A clinician decides
Applications draw a scoped view of the model for their task. The system proposes; a clinician decides.
A context window is the information given to an AI model for one task. A context layer is the persistent model from which many windows can be drawn. One is assembled for an inference; the other is maintained independently of any inference.
Read the whitepaper → · Context layer vs RAG →
The electronic health record is a system of record: it preserves what was decided and documented. A clinical context layer holds how those facts, and signals outside the record, relate to one another over time. It does not replace the record. It draws on it and can feed insight back through alerts and decision support.
No. Clinical Context AI (IMI) is specified, with a prototype next, and is not in clinical use. The patient examples in our articles are synthetic and say so. If you would like to discuss the architecture, talk to the founders.
No. The system proposes and a clinician decides. It never asserts a clinical fact about a patient on its own, and no clinical action is automated. Its job is to keep the relevant context connected and visible so a clinician can judge it.
A context window is the information given to a model for one reasoning task, assembled for that inference and then discarded. A context layer is a persistent model of the relationships among facts, maintained independently of any inference, from which many context windows can be drawn. One is temporary; the other is cumulative.
Last reviewed
The whitepaper: why assembling a context window and maintaining understanding are different responsibilities.
ReferenceShort definitions of the terms used on this page.
ComparisonHow a context layer differs from retrieval, data warehouses, EHR decision support and AI assistants.
Talk to the founders about where the context layer fits in your program.
Talk to the founders
