AI for healthcare teams, with a clinician always in the loop.
Clinics, provider groups and digital health products use us to take administrative load off staff and to evaluate the AI features they are building, without a model ever making a clinical decision.
Where AI earns its place in healthcare.
Patient intake and scheduling agents
Voice and chat agents that follow a clinician-approved protocol to capture details, book and reschedule, and escalate anything urgent or outside the script to staff.
Voice →Referral and clinical-document summarisation
Draft summaries of referrals, discharge letters and histories for a clinician to review and sign, cutting reading time without removing judgement.
Custom AI →Coding and prior-authorisation assistance
Prepare coding suggestions and authorisation paperwork from the record, with sources shown, for the billing team to confirm.
Custom AI →Knowledge assistants for staff
Answers grounded in your protocols, formularies and policies, with citations, for the people who have to apply them.
Custom AI →Evaluation of clinical LLM features
Independent testing of accuracy, safety behaviour and failure modes for AI features in health products, reported for clinical governance and regulatory review.
Evaluation →Fine-tuning for specialty vocabulary
Models adapted to clinical shorthand, specialty terminology and your document formats, trained in your environment on de-identified data.
Training →Designed for the review, not retrofitted for it.
Special-category data
Health data is handled under GDPR and UK GDPR special-category rules, with a DPA, purpose limitation and data kept in your environment.
HIPAA path
We do not describe standard deployments as HIPAA compliant. For US providers we work through a Business Associate Agreement with your chosen vendors and controls.
Not a medical device
Nothing we build diagnoses or recommends treatment. Where a feature could fall under medical-device rules, we say so early and design around it.
Clinician-approved protocols
Any patient-facing agent follows a script your clinical team approves, with fixed escalation rules and full call logs.
Clinical governance evidence
Evaluation reports written so your governance committee can see test design, results and known limitations.
What we will and won’t do in healthcare.
Common questions from healthcare teams.
Can the AI handle urgent or clinical questions?
It follows a protocol your clinicians define: fixed questions, recorded answers, and routing rules such as offering an urgent slot or directing a caller to emergency services with your approved wording. It does not assess urgency itself, and every such interaction is logged.
Is this HIPAA compliant?
Standard deployments are not represented as HIPAA compliant. For US healthcare use cases we discuss whether a compliant path, with the required vendors, Business Associate Agreements and controls, is feasible for your situation before any work starts.
How do you handle patient data in training?
Only with a documented legal basis, de-identification and your data protection team's sign-off. Many training goals can be met with synthetic or de-identified data, and we prefer that route.
Can we start with one clinic or one feature?
Yes. A discovery sprint on one workflow, or an evaluation audit of one AI feature, is the usual first step.
Tell us what you're trying to ship.
Thirty minutes with the founder. We'll tell you honestly whether we're the right team, and what it would cost.