Will the patient know they are talking to a machine?
The agent introduces itself as a call from your hospital. It does not claim to be a nurse or a doctor, and it does not use a person's name.
Which languages does it speak?
More than 18, including the ones your patients actually speak at home. A language is added by a native speaker who has sat in an eye OPD, with your doctor checking the clinical wording. Not by a translation model.
Does the agent diagnose, or give medical advice?
No. It asks the questions your doctor signed, records the answers and applies your tiers. Any reassurance it gives is read from wording you approved.
What happens when an answer is unclear?
It is read as the higher tier, never the lower. The failure mode is a nurse call that was not needed, never a red flag that was missed.
Who does it escalate to?
The nurse, or nobody. There are two routes. The agent never decides which doctor, department or desk a patient belongs to. Your nurse does that.
Do we need to integrate it with our HIS or EMR?
No. It starts from a file your team already exports and gives you a list back. There is no integration project and nothing to install.
Who is responsible for patient data?
You are. Under the DPDP Act 2023 the hospital is the data fiduciary and NexaNeura is the processor. An escalation carries a patient ID, a tier and a callback number. No name, no symptom text.
Is outbound calling to patients allowed?
Yes, with the hospital as the registered principal entity under TCCCPR and every number scrubbed against DND before it is dialled.
What if the patient does not pick up?
It tries again at different times of day, on a schedule you set. A patient who is never reached goes on the nurse's list. Nobody is dropped silently.
How long before it is running?
Front desk goes live in a week and needs no signature. A care pathway needs one hour of a doctor, once, to correct and sign the questions.
