SS
Luxembourg
Siva Subbiah — Founder & CEO
Vehicle dynamics engineer. Background in complex systems modeling applied to healthcare. Leads product strategy, regulatory affairs.
LIET turns doctor-patient conversations into structured clinical notes and longitudinal patient memory — with optional suggestions you can review or ignore. Less time on paperwork. More time with the patient. The doctor is always in control.
Watch what the tool sees as the doctor speaks: live capture, gentle nudges on what hasn't been asked, and patterns surfaced for the doctor's review as they emerge. Shown here: a Tamil–English consultation, among the hardest LIET handles — the same flow runs in French, German, and English.
LIET is built for the multilingual reality of primary care in the Greater Region, where the consultation and the record are often not in the same language.
Not a video — the working prototype, embedded live. Dr. Claire Hoffmann runs a solo cabinet in Belair. Marie Weber, 58, arrives with chest pain on exertion. She speaks French; her note is structured in clinical English as a standards-based record (FHIR R4), designed to align with Luxembourg's DSP and the European EHDS exchange format as those specifications are published. Click through the morning yourself.
Dr. Hoffmann and Marie Weber are fictional. No real patient or physician data appears anywhere on this site.
The doctor's workflow doesn't change. Speak naturally. The system organizes everything.
Consultation happens naturally. Real-time transcription captures every symptom, pertinent negative, examination finding, and medication — organized as it's spoken.
Positive history separated from negative history. Exam findings extracted. Brand names of medications captured accurately. Findings worth a second look are surfaced for the doctor's review. Investigation options listed. All in seconds.
Editable clinical document. Prescription editor. Physician-requested reference tools — clinical evidence, referral templates, investigation checklists. WhatsApp delivery to patient. Everything auditable. The physician confirms every output.
A 45-year-old woman. An 8-minute consultation in a regional Indian language. Here is what LIET captured and organised — without the doctor typing a word.
Illustrative example. The patient and consultation are fictional — a representative scenario built to demonstrate how LIET organizes a complex consultation. No real patient data appears anywhere on this site.
Findings the doctor mentioned but did not dictate into the note were flagged as uncaptured, so nothing said in the room is lost from the record.
Separated positives from pertinent negatives. Extracted the examination. Captured medications with dose and response, and the family history verbatim. Grouped the investigations the doctor dictated by system. The doctor decides everything; LIET writes it down.
4 speech-to-text accuracy errors identified in this scenario (drug name variants, family history formatting).
Why documentation is the entry point — and why the data layer compounds quietly.
Across primary care, the same thing happens every day: the consultation happens, the note gets typed later from memory, and the detail is lost. The pattern across visits, the finding buried in a multilingual conversation — it disappears.
Luxembourg has roughly 778 practising general practitioners, working across four or more languages, under rising administrative load. Almost none of what is said in those consultations becomes structured data.
Europe is mandating structured health data. Luxembourg's CNS-AMMD convention has been digitising administrative transmission since July 2025 — fee statements first, extending to prescriptions and medical reports through 2027. Under the EHDS Regulation, patient summaries and electronic prescriptions must be structured and exchangeable by 2029. The formats are being defined. The capture problem — how structured data gets created during the consultation — is not.
Doctors need their notes done faster — that's the immediate value. But every note generated and reviewed by a physician creates something that doesn't exist elsewhere in primary care, in any market: searchable, longitudinal clinical data in the doctor's own corrected words. The documentation tool is useful on day one. The data layer it builds becomes valuable over time.
Each consultation adds to a growing knowledge substrate — physician-reviewed, multi-language clinical records with corrections, confirmations, and patterns across visits. The longer the system runs in a hospital, the harder it becomes to replicate what it has learned.
What the product does today, and what comes next — we tell it straight. The doctors using LIET today shape the platform that serves the wider clinical community tomorrow.
AI-assisted documentation with suggestions you can review or ignore. LIET captures what's said and organizes it. The doctor decides everything.
When a patient returns months later, LIET surfaces their previous medications, the earlier reasoning, the investigations ordered — without manual chart-pulling. Memory across visits, in the doctor's own corrected words.
Across consultations, LIET surfaces patterns worth a doctor's attention — recurring presentations, response to treatment over time, subtle changes in a patient's trajectory. Brought to the individual visit, only with explicit consent and full de-identification.
Patient privacy and physician control come first, always. No identifiable patient data is ever shared.
A small team building multilingual clinical documentation for primary care.
Vehicle dynamics engineer. Background in complex systems modeling applied to healthcare. Leads product strategy, regulatory affairs.
Trained internal medicine physician. Leads clinical evaluation, product accuracy, and physician partnerships.
Full-stack engineer leading the clinical platform build. Production clinical architecture, encounter flow, and infrastructure reliability.