Initiative overview
The Heidi AI Scribe for Emergency Departments (ED) is an ambient, generative AI clinical documentation resource deployed nationally through the HealthX innovation programme. This AI solution listens to clinician–patient interactions (with appropriate consent), securely records consultations and transcribes in real time. It automatically generates structured clinical notes aligned to documentation standards and clinical terminology. Clinicians review, edit and confirm these notes for the clinical record. Clinical accountability and decision making remain fully with the clinician at all times.
The AI scribe is suitable for high pressure ED environments, where clinicians face significant administrative burden alongside acute clinical decision making. The tool is saving clinician time while making sure important information is accurately captured. Nationwide rollout covered 39 sites with over 1,000 licences allocated. This national deployment across all emergency departments in New Zealand represents a significant milestone in responsible AI adoption at scale.
The deployment is understood to be one of the world’s first fully national deployments of a public health ambient AI scribe for Emergency Departments. The AI scribe enables more direct patient engagement through by freeing clinicians from manual note taking and documentation. Clinicians also report the ability to see additional patients per shift. The solution improves waiting time through improved patient flow and throughput in high demand ED environments, as well as shorter patient stays.
Clinician satisfaction scores exceeded targets, with feedback consistently highlighting reduced cognitive load, improved focus on patient interactions, and less end of shift fatigue. Documentation completed outside rostered hours decreased by up to 81%, reducing unpaid overtime and supporting better roster alignment.
Results, outcomes and impacts
AI scribe improvements are likely to be resulting in a range of positive impacts for patients and their families including reduced waiting time for assessment and faster processing of referrals. Direct improvements measured include time savings. The AI scribe created a substantial reduction in documentation effort and time. Data captured during piloting and early rollout demonstrated reduction in documentation effort of over 10 minutes per patient assessment on average; equal to improvement of approximately 60%. Clinicians reported the ability to see additional patients per shift, driven by faster documentation processes. Documentation completed outside rostered hours also decreased by up to 81%, reducing after‑hours administration.
