In recent years, clinical documentation improvement tools have increasingly become a staple companion for doctors.
This assistive tool, which is helping address clinician burnout, has evolved from just automating transcriptions during clinical encounters or consultations to also being capable of ambiently listening in the background and providing smart recommendations, all powered by AI.
The adoption of smarter, more accurate clinical documentation tools – popularly called digital or AI scribes – has grown over the past two years amid the rave over generative AI.
Now, as AI developers are working on a newer form of AI – agentic, which is capable of acting on its own – it is highly likely that clinical documentation would be one of the emerging AI’s first use applications.
And indeed, a squad of Singaporean doctors is already making this happen.
AIGP Health, a startup founded by four doctors in Singapore, offers an agentic AI-powered clinical assistant that also functions as a real-time copilot supporting clinical judgment. In their words, they said the AI engine, called Anzu, autonomously initiates structured history taking, triage, and patient follow-up before a clinician picks up a case.
The AI assistant, deployed via web and WhatsApp mobile application, passively listens before and during a patient consultation, processes structured and unstructured clinical input, and then generates “accurate, audit-compliant, and context-aware” consultation notes.
Besides improving notes completeness, Anzu reportedly reduces clinical documentation time “by up to 40%” and “surfaces critical patterns that manual workflows often miss.”
A solution to clinicians’ frustrations
In an interview with Mobihealth News, each of the AIGP founders detailed how they started seeking an AI-powered solution to an all-too-familiar struggle among health professionals: time taken away by administrative tasks.
“In emergency settings, every minute matters. Yet, I often spent as much time documenting cases as I did treating them. It’s frustrating when documentation takes precedence over patient presence. I’ve seen how this delay affects triage accuracy, discharge planning, and follow-up adherence. AIGP Health was born from this urgency – to give that time back.”
Dr Nicholas Chia, CEO, emergency medicine specialist
“As a primary care physician, I’d spend 15-20 minutes per patient – of which 30%-40% went to notes, summaries, and follow-up scheduling. That’s time I didn’t have to listen deeply or personalise care. Worse, patients with complex chronic conditions often slipped through. This tool gives us the space to be clinicians again, not just data entry operators.”
Dr Prateet Singh Narula, Medical Lead, primary care and chronic disease specialist
“I’ve worked on the frontline and in AI labs. The disconnect was clear: tools were being built without clinical reality in mind. That’s why our assistant was designed not just to function, but to fit. We focused on creating an AI that doctors trust and that learns from real-life usage. It’s not just about reducing workload; it’s enabling better, faster decisions.”
Dr Yudara Kularathne, CTO, emergency physician and AI Expert
“As a clinician, startup founder, part-time Master of Science student, patient advocate, and parent, I’ve had to adopt an AI-first mindset just to stay afloat. From managing research, team workflows, to even structuring my own learning, AI has become my co-pilot. So when it comes to clinical care, the logic is the same. Patients with autoimmune diseases don’t present with textbook flares; they evolve subtly, unpredictably. If we don’t track that nuance – medical changes, serologies, lived experience – we risk missing what matters. Yet, the mental bandwidth it takes to document all of this, by hand, is no small thing. AIGP wasn’t just born out of convenience. It’s about reclaiming cognitive space, so we can think clinically, act decisively, and focus on what actually changes outcomes: the patient in front of us.”
Dr Anindita Santosa, Product Development, primary care and rheumatology
Not another AI scribe
The market is already inundated with various options for clinical documentation (including those that incorporate AI), with some health systems developing in-house solutions.
“You’re absolutely right. There are many AI scribes and documentation tools entering the market 1750134365. But most of them were built for general-purpose transcription, retrofitted for healthcare, or developed outside clinical environments,” said Dr Narula.
The market for clinical documentation improvement tools could be worth over $10 billion by 2034, rising from around $5 billion today, based on the latest projections. This growth is driven by the increasing volume of patient data healthcare providers collect, compliance with government regulations, and adoption of EMR/EHR systems.
A major player in this space is Nuance, which Microsoft fully acquired in 2022. In Asia-Pacific, Heidi Health from Melbourne, Australia, is growing in popularity; the startup recently raised $17 million in investment to further expand its AI scribe’s capabilities. Meanwhile, big health systems across Asia, including SingHealth in Singapore, South Korea’s largest hospital, Asan Medical Center, and the private healthcare provider group Ramsay Health Care in Australia, have developed their own AI-powered, ambient clinical note-taking tool.
“We’re not building another AI scribe; we’re building a clinical cognition layer,” Dr Narula maintained.
By clinical cognition layer, he means that what they offer is not a passive transcription tool but “an active clinical assistant that thinks in terms of disease trajectories, flare patterns, and decision support surfacing insights that go beyond what was said in the consultation.”
This is unlike existing clinical documentation tools that use generic speech-to-text wrapped in templated SOAP formatting, he added.
AIGP also plans to develop a hybrid architecture that supports on-device edge processing for privacy-sensitive consultations, low latency, and offline resilience – something that cloud-only scribes have yet to do, according to Dr Narula.
Anzu, he further explained, is about “redefining what an assistant can be – an intelligent co-pilot that restores clinical focus, enhances patient continuity, and adapts to real-world healthcare challenges with context, care, and clinical integrity.”
Initially designed for general practitioners in high-volume primary care settings, Anzu can also be used by:
-
primary care physicians and family medicine practitioners;
-
specialists with chronic case loads, such as in endocrinology, rheumatology, geriatrics;
-
nurse practitioners, allied health professionals, and care coordinators;
-
public health teams conducting outreach, screening, or community follow-ups.
It can also be deployed in the following settings:
-
outpatient clinics and polyclinics
-
telemedicine platforms and hybrid digital clinics
-
community health centres and mobile health units
-
home-based chronic care programs
-
post-discharge care coordination units
-
rural health outreach programs
“Because Anzu is designed to be platform-agnostic and accessible via low-barrier interfaces like WhatsApp, it adapts well to both tech-forward urban health systems and resource-constrained environments,” Dr Narula noted. For now, the AI assistant can be integrated into EMR systems via secure APIs while the company works to comply with widely adopted data standards, such as HL7 FHIR.
Additionally, the Anzu co-pilot is one of the few AI-powered tools trained for the Asia-Pacific healthcare environment. It provides multilingual and code-switching support for localised care and tailored documentation styles and alignment with local billing workflows.
Assistive, not authoritative
AIGP explained that its platform is built by taking a privacy-by-design and security-by-default approach, which ensures health data protection at every stage of interaction across platforms or devices.
It features end-to-end encryption, role-based access, and secure audit trails. Connections with external systems are done through secure, authenticated API integrations with strict token-based access and data isolation protocols. It also requires explicit permission to allow third-party systems to access patient data.
Meanwhile, its natural language processing system includes confidence scoring, escalation triggers, and automatic handover to clinicians if ambiguity, risk, or incomplete input is detected
“Anzu never acts autonomously on clinical decisions. All outputs are reviewed, approved, or edited by a clinician before being recorded or actioned, ensuring that AI remains assistive, not authoritative,” Dr Narula stressed.
The AIGP platform is also aligned with Singapore’s Personal Data Protection Act and configurable to comply with other jurisdictions, like the European Union’s (EU) General Data Protection Regulation.
Additionally, the startup will pursue a voluntary SOC 2 cybersecurity certification to “future-proof against scale and policy requirements.”
AIGP is finishing a pilot study with two primary care clinics, with early feedback from clinical staff being “highly positive,” particularly around workflow efficiency and ease of integration.
“Initial results indicate a measurable impact, with clinics reporting at least a 10% reduction in time spent on routine documentation and administrative tasks,” Dr Narula shared.
Approaching change
Pushing the uptake of many digital and technological tools amid digital transformation may spook and overwhelm doctors, especially those who are less digitally savvy.
“We understand that many senior doctors aren’t resistant to innovation – they’re resistant to complexity. Most have experienced tech that over-promised and under-delivered, often adding more steps to their workflow rather than removing them,” noted Dr Narula.
“Doctors don’t need more dashboards—they need more time,” he emphasised.
With their AI-powered clinical co-pilot, AIGP says they take a modular approach to adoption. “Senior doctors can start with one function, like automated history-taking, and gradually expand to clinical summarisation and follow-up tools as comfort builds.”
Anzu is available on WhatsApp, which many clinicians are already familiar with and are using daily. “This immediately lowers the barrier to entry and speeds up onboarding,” Dr Narula claimed.
The assistant also offers transparency: it allows doctors to see, edit, and approve every output before they are saved or sent, “restoring a sense of control rather than forcing a new workflow,” Dr Narula explained.
Moreover, it provides explainable AI outputs, audit trails, and clear documentation protocols.
“Our goal is not to digitise healthcare for its own sake, but to amplify the clinician’s expertise without asking them to change who they are or how they think. The best tech is the kind that quietly supports – without getting in the way.
“We see AI not as an add-on, but as a clinical infrastructure: purpose-built to restore cognitive space, reduce administrative drag, and enable more precise, continuous care.
“Ultimately, our goal is to empower any healthcare provider burdened by documentation and follow-up, and any health system looking to scale personalised care without scaling burnout,” Dr Narula said.
In the next 12 months, AIGP plans to secure the Singapore Health Sciences Authority’s regulatory clearance and conduct final pilots and technical validation with outpatient and GP clinics before commercially launching its agentic AI-powered platform. It will also be preparing for entry into Australia.
The company also shared that in the next three years, it will seek clearance from the Therapeutic Goods Administration of Australia, deploy its AI solution to primary care and telehealth clinics in Singapore and Australia, expand its application to cover more chronic care and documentation use cases, and pursue more secure integrations with third-party EHRs. The company also looks to obtain CE marking from the EU.
AIGP also expects to gradually grow from clinic-level rollouts to health network-level implementations within regulated markets.
