AI assistant for physicians

Just ask Oliver.
With your voice, whatever you need.

Anything irreversible is read back to you and waits for your “yes”.

Operates underHIPAAHL7 FHIRNOM-004NOM-024COFEPRIS
One assistant for your whole day

Everything you do as a physician, ask for it out loud.

It doesn’t just write the note. Oliver helps with every part of your practice: before, during and after each visit.

Visits

Listens, reasons through the case with you and flags what might be slipping past.

“Oliver, start the visit.”

Charts and notes

The note is ready when the door closes, with the patient’s full history at hand.

“Add that she denies fever.”

Prescriptions

Suggested medications with doses calculated from weight and kidney function.

“Amoxicillin for Sofía, 18 kilos.”

Tests

Orders labs and imaging and interprets the results against the case.

“Order a CBC and a metabolic panel.”

Schedule

Ask what’s on tomorrow or open a chart with your voice.

“Move tomorrow’s appointments to the afternoon.”

Referrals

Drafts the referral to the specialist with a summary of the case.

“Refer him to cardiology with the summary.”

Insurance

Fills in the insurer’s forms from the same chart.

“Bill this visit to the insurer.”

Follow-up

Treatment adherence, pending items, results and alerts for every patient, without relying on memory.

“Follow up with Juan on his treatment.”

This is what it looks like in a real visit.

Try it yourself. One visit, start to finish.

Start the visit, follow the guided exam on the body and watch the note come out. Sample data, no sign-up.

Oliver
LRLucía Ramos

00:00

Listening…

…my head has been hurting for three days

Clinical decision support

Clinical answers, without leaving the visit.

Ask out loud or in writing. Oliver answers with the visit as context and every claim linked to its source: guidelines, literature and the chart itself.

Visit in progress · 08:42
Example
Should I adjust metformin with this kidney function?

With an eGFR of 38 mL/min, the guideline recommends reducing the dose to a maximum of 1,000 mg a day and monitoring kidney function every 3 to 6 monthsKDIGO 2022 · Diabetes in CKD. In the chart, eGFR was 52 eight months ago: the decline is recentLabs · February.

Add to planView sources

What might be slipping past, in time.

While you talk with the patient, Oliver cross-checks what it hears against their history and flags red flags and the questions that would change your management.

How it works
Listening · example
Red flag
Exertional chest pain with family history of heart disease: rule out acute coronary syndrome.
Suggested question
Does the pain radiate to the arm or jaw?
Live red flags

They appear quietly while you see the patient. If they don’t apply, dismiss them with a tap.

Questions that change management

Only the ones that would move your diagnosis or plan. Not a questionnaire.

With the whole history

Prior visits, tests and current medications feed the reasoning without you searching for them.

Arrive prepared. See patients without a keyboard.

Pre-visit
Reason for visit captured
History updated
Pre-diagnosis ready to review
The patient arrives with context

Before the visit, Oliver gathers the reason, history and a first line of reasoning so you walk in knowing why.

History · step 3 of 7
When did the pain start?
HoursDaysWeeks
Guided history-taking

When you want it, Oliver walks you step by step through the history for that complaint, leaving no gaps.

“Oliver, add that she denies fever.”
Subjective · Denies fever.
Hands-free, with your voice

Ask for changes, open a chart or start the visit by speaking. Anything irreversible is confirmed first.

You close the door and it’s all done.

The note, prescription, test orders and diagnosis codes all come from the same visit. Oliver proposes; nothing is signed without you.

See how signing works
To sign · example
Sign all
Clinical noteSOAP format
Metformin 500 mg every 12 hAdjusted to eGFR 38
Serum creatinine · HbA1cLab order
E11.2 · Type 2 diabetes with renal complicationsN18.3 · Chronic kidney disease, stage 3
The plan as a cart

Oliver recommends medications, tests and instructions. You choose what goes in and sign once.

Doses calculated, not guessed

Doses come from a deterministic calculation using the patient’s weight and kidney function, not the language model.

Codes with their rationale

Each ICD-10 code arrives with the part of the visit that supports it.

Ready for the insurer

If the patient is insured, the forms are filled from the same chart.

The visit ends.
Oliver stays on the case.

When results come in, Oliver reads them against the case and tells you what needs action. Each patient’s pending items stay in view, not in your memory.

Results interpreted

Labs and imaging read in the context of the visit that ordered them.

Pending items per patient

Who is waiting on tests, on an appointment, on an answer from you.

Your day at a glance

Ask “what’s on tomorrow?” and get a summary.

Integrated with Colegas

Your network of peers, inside your assistant.

Oliver integrates with Colegas, the professional network for physicians: share clinical content, discuss cases and join networks in your specialty without leaving where you work.

Discover Colegas
Colegas · example
Cardiology · Mexico CityCardiology network

De-identified case: atypical chest pain in a 52-year-old woman with an inconclusive stress test. Coronary CT angiography or stress echocardiogram?

Shared from the visit · patient data de-identified

CardiologyPediatricsInternal medicineGynecology
Share clinical content

Post cases, guidelines and lessons. Patient data is de-identified before it leaves your chart.

Specialty networks

Join groups of peers in your specialty and follow what they discuss.

Peer consults

Ask a colleague for a second opinion and get the answer right next to the case.

Trust and compliance.

A clinical assistant is only useful if the physician can audit what it proposes and the patient can trust where their data lives.

Trust center
Built for NOM-004 and NOM-024

A structured clinical record that follows Mexican regulation and interoperates over HL7 FHIR.

Patient data encrypted

Clinical information is stored encrypted and every access is logged.

Every suggestion with its source

See where each claim comes from before you accept it.

The physician decides and signs

Oliver proposes. Nothing enters the chart or reaches the patient without your signature.

You speak.

Oliver does it.

Writes the note.

Your AI assistant, built on HL7 FHIR.

Start free. Nothing to install.

Anatomical model: BodyParts3D, © The Database Center for Life Science, adapted. License CC BY 4.0.