The world’s clinical AI is being built on a narrow slice of medicine
Almost all the data behind today’s clinical AI comes from the United States and Western Europe, in English. Yet these systems are being deployed globally — into different disease patterns, different resource realities, different clinical norms and different languages.
We work with healthcare organisations around the world to license real-world clinical data for AI training and evaluation, and we do it in a way that holds up to scrutiny.
- Regions
Europe · Africa · Asia · Middle East
Providers, clinics, digital health, research
- De-identification
In-jurisdiction
Identifiers removed before data leaves
- Licensing
Non-exclusive
By default
- Evidence
Published research
Clinical AI evaluation (arXiv)
Who we work with
Designed for both sides
You already hold clinical records. We turn them into a revenue line without changing how you deliver care, and without your patients’ privacy being the price. We handle buyer relationships, specification, quality assurance and delivery. You stay in control of your data and your patients’ trust.
Consent and privacy
Identifiers are removed before data ever leaves your jurisdiction. We work with you on lawful basis, consent and de-identification, and nothing is licensed exclusively unless you choose it.
We source clinical data across languages, specialties and health systems that are largely absent from existing corpora — consultations, clinical records, imaging, structured and unstructured data. Every source is assessed on provenance, rights and fitness for purpose before we put it in front of you.
Provenance and rights
Every source is assessed on provenance, rights and fitness for purpose before we put it in front of you.
What we license
Requirements vary by buyer. What stays constant is that the data is real, lawfully obtained, and fit for the purpose it is bought for.
If you hold something that isn’t on this list, it is still worth a conversation. Buyer requirements change constantly, and we source to specification.
- Patient–clinician consultations — transcripts, audio and video
- Clinical notes, records and structured EMR data
- Specialty and allied health data, including diagnostics and rehabilitation
- Clinician-authored evaluation and benchmarking data
- Model outputs paired with real clinical decisions
How it works
How a partnership works
We establish what you hold, in what languages and volumes, and who controls it. No commitment at this stage.
We work with you on lawful basis, consent and de-identification — designed so that identifiers are removed before data ever leaves your jurisdiction.
Per-unit licensing, structured as an initial tranche plus recurring volume. Non-exclusive by default.
We manage specification, quality assurance and the buyer relationship. You supply data; we do the rest.
Healthcare AI evaluation
We find where healthcare AI breaks
Your model passes benchmarks. We test if it’s safe for patients.
We design and produce healthcare datasets that test how models behave under real clinical conditions — including uncertainty, incomplete information, and evolving patient states.
Our focus is not on what models know, but how they reason.
Failure-driven design
We target scenarios where models produce plausible but incorrect decisions.
Real clinical complexity
Cases include ambiguity, conflicting signals, and time-sensitive decision-making.
Structured evaluation
Each task includes expert-built rubrics focused on reasoning quality and safety.
Quality above all
As models improve, the bottleneck is no longer data volume — it’s whether the data actually exposes where models fail.
We design datasets that go beyond textbook scenarios and pattern matching. Our focus is on:
- real-world clinical ambiguity
- incomplete and conflicting information
- edge cases and failure modes
Every line of data is built to test reasoning, not recall — and ultimately answer one question:
Does this make the model better?
Deep experience in healthtech & AI
We’ve spent over 6 years working at the intersection of healthcare and AI.
This includes:
- supplying clinicians into healthtech and AI environments
- working closely with how models are trained, evaluated, and improved
- understanding where models succeed — and more importantly, where they break
We don’t approach this as a data vendor, but as a partner focused on improving real-world model performance.
10,000+ clinician talent pool
We have built a network of over 10,000 clinicians, including specialists across key domains such as oncology, cardiology, and paediatrics.
Many of our clinicians:
- have experience working with frontier AI systems
- understand evaluation frameworks and model behaviour
- can contribute beyond annotation into reasoning, critique, and refinement
This allows us to deliver high-quality, expert-driven data at speed — without compromising on depth.
Two people, one on each side of the problem
Nurture AI is run by its two co-founders. One has spent years working with clinicians and the organisations that employ them. The other has spent years building data systems in institutions where governance and audit are non-negotiable. Between them is the whole of what a data partnership needs: the clinical relationships, and the technical discipline to handle what those relationships produce.
Ali Merali
Co-founder & CEO
Ali leads Nurture AI’s partnerships with healthcare organisations and AI labs, and runs its clinical evaluation programme. He is the founder of MAKZ Talent, where he has spent more than six years placing clinicians into healthtech and AI environments and has built a network of over 10,000 clinicians, including specialists in oncology, cardiology and paediatrics. That work gave him a working understanding of how models are trained, evaluated and improved — and where they break in real clinical settings. He works between London and Florida.
LinkedIn — Ali MeraliMurtaza Merali
Co-founder & CTO
Murtaza is responsible for how data moves through Nurture AI: de-identification, quality assurance, secure delivery and the tooling behind each partnership. He has over twelve years in data and analytics, including as a Vice President at J.P. Morgan and earlier roles at Barclays and Santander — environments where data access, lineage and audit are treated as seriously as clinical data should be. Since then he has built data platforms on Google Cloud for financial services and retail clients as an independent consultant. He holds a degree in mathematics and works between London and Muscat.
LinkedIn — Murtaza MeraliWhen you work with Nurture AI you deal with the founders directly. There is no account team between you and the people making the decisions.
If you hold clinical data, you may be holding something valuable
We work with healthcare organisations across Europe, Africa, Asia and the Middle East — providers, clinic groups, digital health platforms and research organisations.
We are particularly interested in data from languages and health systems under-represented in existing training data.
If you would like to understand what your data might be worth, and what a responsible arrangement would look like, we would welcome a conversation.
Backed by published research on clinical AI evaluation. Read the paper