What Is an AI Co-Clinician?

The term is new and the job is not. An AI co-clinician reaches patients, gathers information and prepares options, and a clinician makes every decision. Here is the longer version of the definition on our Gia page.

What is an AI co-clinician?

Three words in that definition carry the weight. Reaches: the co-clinician initiates contact, between visits, with people the team cannot get to. Prepares: it turns what it gathered into a ranked list with reasons and options. Decides: it never does. That verb belongs to the clinician.

The category exists because the care team's limiting factor is not knowledge. It is hours. A nurse knows which patients need a call this week. She does not have the week to make them.

An AI co-clinician is clinical decision support, not a diagnosis, and it is not a replacement for anyone on the team. It is the member of the team whose job is the part nobody else has time for.

How it differs from a chatbot

A chatbot waits to be asked. A patient opens an app, types a question and gets an answer. Nobody on the care team sees the exchange, nothing is written to the record, and the patient who never opens the app is invisible.

An AI co-clinician calls. It does not wait for the patient to come to it. When nobody answers, it retries. What it learns goes to a clinician, not back to the patient as advice, and the clinician's decision goes to the record. The direction of the conversation is reversed, and so is the accountability.

How it differs from an ambient scribe

An ambient scribe listens to a visit that is already happening and writes the note. It is useful, and it is about documentation. It does not reach anyone, and it has nothing to say about the patient who was not in the room.

An AI co-clinician works between visits, with the patients who are not in the room. The two tools answer different questions: the scribe asks how the clinician documents the visit; the co-clinician asks which patients should have one. Many teams will end up with both, and the scribe will document the visits the co-clinician made happen.

How it differs from a risk score

A risk score is a number from the record. It tells a team which patients to worry about, and then it stops. Someone still has to make the call, and the score does not know that the patient's speech has changed since the last visit, because speech is not in the record.

An AI co-clinician acts on the list. Scienza's foundation model for aging reads the records a plan or system already holds and reports which cohorts carry expected risk, counted, never named. Gia then reaches those people and adds what the record lacks: a 40-second speech check-in measuring 2,500 speech biomarkers, pace, pauses, articulation and prosody, with results to the clinician in under two minutes. The score says where to look. The co-clinician goes and looks.

Who deploys one, and what the clinician keeps

Health plans, ACOs, health systems and senior-care operators deploy Gia to check in with older adults between visits, by phone or video, in 44 languages. Gia is a Digital Human®: a face and a voice on video for those who want one, a voice on the phone for everyone else, including a landline.

The clinician keeps every decision. Each result arrives as a ranked worklist with the reason beside each name and Proactive Decision Actions. Options, never orders. The clinician chooses a visit, a nurse call, a medication review, a structured assessment administered in conversation, or nothing yet. The choice writes back to Epic, PointClickCare or your EHR, dated and attributed.

The clinician also keeps the encounter. Gia does not document a condition, does not perform a plan's health risk assessment and does not tell a patient what she has. Gia brings the clinician the patient, the information and the options. What happens next is medicine, and medicine stays with the clinician.

Where to start

Pick one population your team cannot reach today: patients in the two weeks after discharge, or members over 75 with no cognitive assessment on file. Have Gia make the check-ins for 30 days and route every result to the clinicians already responsible for those patients. Count reach, decisions recorded and visits that happened because of a call.

See how ACOs and health systems deploy an AI co-clinician, then book a working session.