Industry 01
Healthcare
Doctors get more time for their patients, and patients face shorter waits and fewer long trips.
Vietnam's top referral hospitals are overloaded all year. Some handle eight or nine thousand visits a day; some have added clinic hours until 9 p.m. Many patients come from distant provinces on overnight buses, renting rooms and bringing family along, sometimes just to be seen and pick up a prescription. Inside the hospital, doctors examine patients while typing notes and read scans while drafting reports, and electronic medical records are now mandatory at every healthcare facility. New Era works with partner hospitals to bring AI into each department for the repetitive work: listening and taking notes, screening results ahead of the doctor, sending reminders and alerts. The time saved goes back to patients. Every clinical conclusion is still made by a doctor.
Industry challenges
What is holding healthcare back
Every pain point comes with the answer New Era proposes.
-
01 · Challenge
Doctors split between patients and paperwork
Each visit lasts only a few minutes, and a large share of that goes to the keyboard and forms. The pressure lands on the people with the scarcest expertise.
Solution
Voice-based exams and history taking: AI listens and drafts the record, the doctor reviews and signs.
-
02 · Challenge
Scans and lab results pile up
Imaging and lab volumes grow faster than the number of specialists. An urgent case can sit among hundreds of routine ones, waiting its turn to be read.
Solution
AI screens each case first, marks suspicious areas and moves abnormal cases to the top; the specialist makes the call.
-
03 · Challenge
Medication errors and missed signs of decline
A missed allergy, drugs that interact, vital signs slipping overnight. Many errors and complications show warning signs first; there is simply no one free to catch them in time.
Solution
Alerts at the moment of prescribing, plus continuous monitoring that warns the doctor on duty early.
-
04 · Challenge
Patients travel far to see a good doctor
Local facilities lack specialists. Patients travel hundreds of kilometers to higher-level hospitals, paying for transport and lodging while family members lose workdays. Central hospitals fill up with cases that could have been treated close to home.
Solution
Clinics in a box: patients are seen near home, and a specialist at a higher-level hospital examines remotely and signs off on the final decision.
-
05 · Challenge
No room for error in health data
A data error, or a conclusion a machine reaches on its own, can cost someone their health.
Solution
AI only suggests. The doctor decides, and every approval is logged. Patients decide who their data is shared with.
Solutions
Solutions for healthcare
Doctors look at patients instead of keyboards; patients spend less time in line at each counter.
Voice-based exams and history taking
Doctors talk with patients as they normally would. AI listens and records the reason for the visit, present illness, medical history and allergies as a draft record in the hospital's own format; the doctor reviews, edits and only then signs.
Intake before the consultation
While waiting, patients describe their symptoms out loud in everyday Vietnamese, so older people who are not used to typing can still give a full account. The doctor gets a summary before the patient walks in.
Queues and routes through the hospital
Patients know in advance when they will be called and the order of the lab and imaging rooms they need. Turns go to wherever is least busy, so no one spends half a day waiting in a hallway.
AI reads first and marks what needs a closer look; the radiologist still signs the report.
AI reads every scan first
X-ray, CT, MRI and mammography images are read first by AI, which outlines suspected lesions for the doctor to examine closely. The doctor still reads the full image and draws the conclusion.
Urgent cases read first
Scans with danger signs such as brain hemorrhage or pneumothorax move to the top of the reading list instead of waiting their turn, and the doctor on duty is alerted at once.
Reports drafted from the doctor's dictation
The doctor dictates the findings and AI drafts the report in the department's template. The doctor edits and signs, and the patient gets results sooner.
Abnormal results reach the treating doctor sooner; small lesions are harder to miss.
Instant alerts for critical results
Lab results beyond critical thresholds go straight to the treating doctor and the nurse in charge, alongside the previous result for comparison.
AI pre-screening in histopathology
Digitized slides are screened by AI first, with suspected malignant areas marked; the pathologist reviews them and makes the diagnosis.
AI-assisted endoscopy
During the procedure, AI flags areas that may hold polyps or early lesions so the doctor can stop and look closely; images are attached to the report automatically.
Critical patients are spotted early; inpatients stay monitored even when few staff are around.
Emergency triage
Vital signs and reported symptoms are checked against triage criteria; patients with critical signs move to the top and staff are alerted at once. The triage nurse still decides the priority level.
Early warning of deterioration
Pulse, blood pressure, breathing rate and lab results for inpatients are tracked continuously. When several indicators worsen together, the doctor on duty is alerted early, even at midnight.
Progress summaries and shift handover
Several days of treatment are summarized for the morning briefing and the next shift, so whoever takes over knows right away which patients need attention.
Medication errors are stopped at the moment of prescribing; records are complete when the patient is discharged.
Safe prescribing
Every medication order is checked against allergy history, drug interactions and dosing by age, weight and kidney function. Warnings appear while the order is written, before it reaches the pharmacy.
Complete records, correct codes
AI suggests ICD-10 codes from the diagnosis and checks the record before it goes to health insurance review; gaps and errors are flagged while the doctor still remembers the case.
Discharge and home care
Discharge papers and medication instructions are drafted in plain language. At home, patients get medication and follow-up reminders, and warning signs are passed to the doctor in charge.
Patients get connected care, from the first booking to every follow-up visit.
Booking and patient care
Patients book on their phones, get reminders on time and can have questions answered after hours. They know their health insurance costs and care plan in advance.
Front desk and scheduling
The whole day's schedule on one screen; walk-ins and booked patients are both checked in quickly. Appointment reminders run around the clock.
Doctor portal and clinic hours
Doctors open their own clinic hours and book on behalf of the patients they refer, without messaging back and forth with the front desk.
Accurate diagnosis
AI and doctors: what the major studies show
Figures from published research by Stanford, Harvard, Johns Hopkins and the world's leading medical journals.
What do patients care about most? An accurate diagnosis.
If I had cancer, what I would care about is a correct, accurate result. Whether it came from AI or a human being wouldn't matter.
The cost of a wrong diagnosis
795,000
Americans die or are permanently disabled each year because of diagnostic errors.
Johns Hopkins · BMJ Quality & Safety, 2023-
Imaging · Chest X-ray
1.5 minutes
for AI to read 420 X-rays; radiologists needed 240 minutes on average
Across 14 common conditions on chest X-rays, the AI matched radiologists on 10, did better on 1 and still trailed on 3.
Stanford · PLOS Medicine, 2018 -
Imaging · Mammography
−9.4%
missed breast cancers in screening, compared with the first reader (US data)
False alarms also fell by 5.7%. Compared with six radiologists, the AI outperformed all six. In the UK, with AI as the second reader, the second reader's workload dropped by 88%.
Northwestern Medicine, Imperial College London · Nature, 2020 -
Pathology · Lymph node metastases
99.5%
accuracy when pathologists and AI read together
Finding metastatic breast cancer cells in lymph node slides, AI was right 92% of the time and pathologists 96%. When pathologists added the AI's findings, accuracy rose to 99.5%, higher than either one working alone.
Harvard Medical School, Beth Israel Deaconess · 2016 -
Dermatology · Skin cancer
On par with 21
board-certified dermatologists at classifying skin cancer from photographs
Trained on about 130,000 images of skin lesions, the AI told skin cancer apart from benign lesions at the level of specialists.
Stanford · Nature, 2017 -
Cardiology · ECG
0.837
average accuracy score for the AI, versus 0.780 for cardiologists
Recognizing 12 rhythm classes, including arrhythmias, on continuous ECG recordings: 91,232 recordings from 53,549 patients.
Stanford · Nature Medicine, 2019 -
Emergency care · Sepsis
14.6%
mortality when a doctor confirmed the AI alert within 3 hours, versus 19.2% when it was not confirmed in time
A study of 590,736 patients at 5 hospitals, used by more than 4,000 doctors and nurses. The AI caught early signs of sepsis, but alerts only helped when a doctor reviewed, confirmed and treated in time.
Johns Hopkins · Nature Medicine, 2022
AI does not replace doctors
The studies point to the same conclusion: the best results come when AI and doctors work together. AI reads fast, never tires and flags what is easy to overlook. Doctors understand the patient, weigh the full history and take responsibility for every conclusion. That is why, in every New Era healthcare solution, a doctor signs off on every final decision.
Figures are taken from published studies. Before deployment, each AI tool is evaluated again on the partner hospital's own data before it goes into use.
Dedicated solution
Clinics in a box
A complete clinic set up at the local level and linked to specialists at higher-level hospitals: patients are seen near home, and central hospitals carry less of the load.
Principle
A doctor signs off on every final decision
AI takes the history, keeps notes, reads results first and makes suggestions. Every diagnosis, every prescription and every referral takes effect only after a doctor has reviewed and signed it, and the system records who approved it and when.
The problem
Patients from the provinces account for 50–70% of visits at the top referral hospitals in Ho Chi Minh City. About 80% of cases that reach central hospitals are mild and could have been treated locally, but local facilities lack specialists. So patients still travel far, queue alongside serious cases and push hospitals past capacity. Every trip means bus fares, lodging and lost workdays for the family, sometimes just to be seen and pick up a prescription.
What is Clinics in a box
Clinics in a box is a complete clinic set up at a commune health station, an area clinic or a community far from the city. Patients give their history, have vital signs taken and get an ECG, ultrasound and rapid tests on site. AI prepares the record and reads the results first; a specialist at a higher-level hospital examines remotely, makes the diagnosis and prescribes.
Key functions
- Voice-based intake. Patients describe their symptoms in everyday Vietnamese, and older people can give a full account too.
- On-site measurements and rapid tests. Vital signs, ECG, ultrasound and rapid test results go straight into the record, with nothing copied by hand.
- AI prepares the record. The history is summarized, results are screened in advance and abnormal findings are flagged for the doctor to look at first.
- Remote exams by specialists. Specialists examine patients by video, see the full record and call in colleagues for a case conference when needed.
- Automated medicine cabinet for signed prescriptions. An on-site cabinet hands out medicine right after the visit, and only for prescriptions a doctor has approved and signed. Without the doctor's signature, the cabinet dispenses nothing.
- Health records on blockchain. Every visit, every result and every doctor's approval is recorded in sequence, and no one can edit or delete anything without leaving a trace. Patients carry their records with them and decide which facilities can see them.
- Follow-up and referral. Follow-up reminders go out automatically; cases that need a higher-level hospital are referred with the full record and an appointment.
Benefits
- Less load on central hospitals Routine cases and chronic conditions that need regular follow-up are handled on site, so higher-level hospitals can keep their beds and time for serious and complex cases.
- Lower travel and lodging costs for patients No bus to the city, no rented room while waiting to be seen, no family member taking time off to come along. Patients go home the same day.
- Top doctors reach remote areas One specialist can see patients in several locations in a single session without leaving the hospital.
- Serious cases referred at the right time Danger signs are caught early. Patients referred to a higher-level hospital arrive with an appointment and their test results in hand.
- One record for the whole course of care The on-site visit, the remote consultation and treatment at a higher-level hospital all sit in one record. The receiving hospital gets complete, reliable information, and patients do not have to repeat their history or redo tests they already had.
Built for
- Hospitals and departments
- Health stations and local clinics
- Clinic owners
- Doctors and nurses
- Front desk staff
- Patients
Results
- Doctors gain more time for their patients
- Urgent cases are read first, and medication errors are caught the moment they are prescribed
- Patients far from the city are seen near home, and higher-level hospitals carry less load
- Patients come back on schedule, and clinics have no empty slots
Projects
Related projects
-
Healthcare
Patient booking and care network
The problem: Every clinic knows the pattern: a last-minute cancellation leaves an empty slot, the front desk calls patient after patient to fill it, and nobody reminds patients when a follow-up is due. Empty slots are lost revenue, and patients miss their treatment.
- Fill cancellations and turn empty slots into revenue
- Fewer no-shows
- Ease the peak-hour crush
-
Healthcare
Clinic operations system
The problem: Appointments are scattered across notebooks, spreadsheets and text messages; referring doctors must message the front desk back and forth to book a slot.
- Keep every referral from partner doctors
- All bookings in one place
- Doctors own their hours
Facing one of these challenges?
Tell us about it. A specialist will get back to you during business hours.