Care planning is often understood as a discrete activity. A plan is created, documented, and then followed.
But for individuals living with chronic conditions, care is not delivered in discrete moments. It unfolds day by day, shaped by small actions and subtle changes. What matters is not only what the plan says, but how it is carried forward over time.
Planning AI extends care planning into this daily context. It determines what actions are appropriate now, given what has recently occurred and what is likely to happen next.
To do this, it must operate on a continuously updated understanding of the individual.
This is the concept of health state.
Health state is not a static record. It is a representation of the individual at a point in time, informed by multiple dimensions: clinical measurements, symptoms, adherence patterns, functional status, and context. It reflects both what has been observed and how those observations are interpreted together.
For Planning AI, this representation is foundational.
Without it, decisions are made in isolation. A single data point may trigger a response, but without context, that response may be incomplete or misaligned. A slightly elevated measurement, a missed action, or a reported symptom may or may not be meaningful, depending on what has come before.
Planning requires continuity.
Planning AI uses the current health state to determine what matters today. It evaluates whether the individual is stable, improving, or beginning to drift. It identifies which aspects of the plan require reinforcement, which require adjustment, and which can remain unchanged.
From this, it generates actions.
These actions are not generic. They are specific to the individual’s current condition and trajectory. They may involve reinforcing adherence, prompting monitoring, adjusting behaviors, or escalating attention when needed.
As new information arrives, the state is updated. The interpretation changes, and so does the plan.
In this way, planning becomes a continuous process rather than a static instruction set.
The quality of this process depends on the quality of the underlying representation. If the health state is incomplete or outdated, planning will be misaligned. If the state is continuously maintained and coherently interpreted, planning can remain aligned with the individual’s evolving condition.
Planning AI, therefore, does not operate on data alone. It operates on a model of the individual that exists over time.
That model is what allows care to be guided continuously, rather than episodically.