INDACO.ONE / Knowledge Centre / Future AI Agents in Care

Future

Future AI Agents in Care

AI agents in care are software systems that carry out multi-step tasks on behalf of care teams — retrieving, summarising, coordinating — within defined limits and under human oversight.

The plausible near term

The realistic contribution of agents in long-term care is coordination rather than judgment: assembling what is known about a person across sources, preparing a summary before a review, drafting the routine documentation that consumes so much of a care team's day, and flagging that something appears to have changed.

These are consequential tasks precisely because they are unglamorous. Administrative burden is one of the largest, and least contested, drains on care capacity.

Where the line must hold

An agent that acts on health information touches decisions that affect people. The boundaries are therefore not optional: a human remains accountable; actions are traceable; the system's confidence and its limits are legible to the person relying on it; and clinical judgment is never delegated.

Under the EU AI Act, systems used in health contexts face requirements around transparency, human oversight and risk management. Designing for those obligations from the beginning is easier and more honest than retrofitting them.

What an agent-ready architecture needs

Agents are only as useful as the substrate they act on. Without a shared data model, an agent has nothing coherent to reason over; without a continuous timeline, it can summarise a moment but not a trajectory; without an API layer, it cannot act; without provenance and auditability, it cannot be trusted.

In other words, the architecture that makes connected care possible today is the same architecture that makes agents possible tomorrow. This is not a coincidence — it is the reason to build the layer before the agents.

How INDACO.ONE approaches it

Future AI agents appear on the INDACO.ONE roadmap deliberately, and deliberately not yet. The platform is being built so that when they arrive they operate on one timeline, through one API layer, with provenance and human oversight already in place — rather than as a feature bolted onto a system that was never designed to be reasoned over.

Related concepts

INDACO.ONE is a software-only Connected Care Intelligence Platform. It is not a diagnostic device and does not diagnose, treat or replace clinical judgment. It supports understanding, coordination and longitudinal insight under human oversight.