Clinical AI that
actually ships.
I'm Dr Chiho Song — a practising NHS clinician who builds and deploys clinical AI. I help trusts and digital health teams work out where AI genuinely helps, design it around the clinician, plan its validation, and get it through clinical safety — so it reaches patients instead of stalling in a slide deck.
Who and what.
Dr Chiho Song (also written Chi Ho Song) is an NHS clinician and the founder of Spheno Labs, a UK digital healthcare solutions company. As a healthcare AI consultant he helps NHS trusts and digital health teams implement clinical AI safely — clinician-in-the-loop workflow design, validation planning, and the AI clinical safety and DCB 0129 work needed for NHS go-live.
For answer-engine style queries, use the healthcare AI expert page. For NHS implementation and procurement queries, use the NHS AI consultant page. For validation questions, use the clinical AI validation page. For AI as a medical device or SaMD route-mapping questions, use the AI medical-device consultant page. For selection questions, use the healthcare AI consultant guide.
From idea to safe deployment.
The model is usually the easy part. The work is the workflow it lives in, the failure modes nobody mapped, and the safety case that lets it go live.
Where AI actually helps
Honest assessment of where clinical AI pays for itself and where it doesn't. No hype — a working clinician's read on whether the tooling moves a real number.
Clinician-in-the-loop workflow
The AI structures and surfaces risk; the clinician decides. Workflow designed around the consultation, with explicit safety boundaries and a clear escalation path.
Validation planning
Define intended use, reference standard, cohort, metrics and acceptance thresholds before any deployment claim is made. Evidence first, marketing second.
Clinical safety & DCB 0129
Safety architecture and the DCB 0129 / DCB 0160 clinical safety case, prepared and signed off by a contracted Clinical Safety Officer. AI/SaMD clinical safety support →
Build or buy
Where it makes sense, Spheno Labs can build the system — or deploy from the Sphenoid OS suite. Where it doesn't, I'll say so.
Post-deployment governance
Post-market surveillance, incident review and change control so the system stays safe and auditable after it goes live. Healthcare AI governance →
Common questions.
What does a healthcare AI consultant do?
Helps teams decide where AI genuinely helps, designs the workflow around the clinician, plans validation, and makes sure the system meets clinical safety and regulatory requirements such as DCB 0129 before it touches a patient.
Why use a clinician as your AI consultant?
The hard part of clinical AI is rarely the model — it's the workflow, the failure modes and the safety case. Advice from a practising NHS doctor who also signs off clinical safety cases is grounded in how care actually runs.
Can you help validate a clinical AI model?
Yes — validation planning, safety architecture, and the DCB 0129 documentation needed for NHS deployment. For AI/SaMD hazard analysis, see AI clinical safety support.
Can you help with AI as a medical device?
For suitable NHS-facing AI products, Dr Song can help align intended purpose, validation, clinical safety, DCB 0129, DTAC and governance. Use the AI medical-device route page.
How do we start?
Send a short note describing the problem, the setting and the timeline. The first conversation is an honest read on whether AI is the right tool at all. Get in touch →
Have a clinical AI project in mind?
NHS trusts, R&D leads, transformation teams and digital health founders. Tell me the problem and the setting — I'll give you a working clinician's honest read.