The principle: AI assists, the doctor decides
Every technical output at HealTrip supports preparation and organisation. It is not a diagnosis or a prescription. Medical decisions are approved by a named, licensed doctor, and you can always ask for human review.
Platform layers
| Layer | What it does | Status |
|---|---|---|
| 1. Consent and identity | Who sees what, with which consent and for how long. | In development |
| 2. Organised health record | Reports, tests and imaging in one chronological file. | In development |
| 3. AI-assisted preparation | Summarising the file and highlighting open questions to save doctors' time. | In development |
| 4. Decision Room | Independent specialist review and a documented decision. | Illustrative demo |
| 5. Care coordination | Next steps with partners: labs, imaging, home care, pharmacies. | On the roadmap |
Research engines
A research engine for reading health signals in context: wearable data, diabetes, blood pressure and heart health, together with environment, season and lifestyle, with modes for Ramadan, Hajj, prayer times, women's health and Saudi cities. The aim: clearer summaries for care teams and early flags that a doctor reviews.
The security and governance layer: encryption, consent-based permissions, an audit trail for every access and decision, monitoring of model risk and bias, and alignment with the Saudi Personal Data Protection Law.
The "evidence before answers" layer: checks AI answers before they are shown, links each statement to its source, prefers "I don't know" to guessing, and refers to a doctor when in doubt.
What these engines do not do
- They do not diagnose, prescribe or change doses.
- They are not an approved medical device and will not be used clinically before clinical validation and regulatory approval, including by the SFDA where software-as-a-medical-device rules apply.
- They are not an emergency service: in an emergency call 997.
- We do not use your health data to train models without explicit, separate consent.
How we will validate before use
- Design review with a medical advisory board.
- Testing on de-identified or synthetic data, measuring errors and bias.
- A limited, doctor-supervised pilot where every output is shown to a doctor before any use.
- Publishing results and limits here before any accuracy claim.
More: Trust center · Launch & service status · Proof of progress