AI Health Dialogue
When AI monitors your health, both the patient and the clinician need to understand what it's saying — and why.
The Current Problem
AI health monitoring platforms collect rich patient data and generate clinical insights — but the explanations stop at the dashboard. Patients receive scores without meaning. Clinicians receive AI recommendations without reasoning. The result is a consultation where neither party fully understands the information between them.
Patients receive scores, not meaning
“Your pain score is 34” tells a patient nothing about what to do next
Clinicians trust the output, not the reasoning
AI recommendations arrive without explanation, making it hard to discuss with patients or challenge when wrong
Patient-reported data gets lost in aggregation
Quality-of-life questionnaires are flattened into numbers before the clinician ever sees the nuance
No audit trail for AI-influenced decisions
When AI shapes a diagnosis or treatment path, neither patient nor clinician has a record of what the AI concluded and why
Our Design Approach
We've designed a bidirectional explainability layer for health AI: interfaces that translate AI findings into language patients can act on, and that surface patient-reported experience in ways clinicians can trust and use. Transparency flows both ways.
Plain-language interpretation
AI findings translated into what they mean for daily life, with suggested questions to ask the clinician
Visible reasoning
Each AI recommendation shows the data points that drove it, so clinicians can validate, override, and explain
Narrative surfacing
AI identifies the most significant patient-reported signals and presents them as clinical talking points before the appointment
EU AI Act Article 13 compliance
Full transparency logs, human override records, and patient-accessible decision summaries
Interface Design Proposals
These mockups show how explainability can be designed into both sides of the health AI consultation. Developed in context of EU AI Act Article 13 transparency requirements and GDPR health data obligations.
Your Health Summary
Based on your last 4 weeks of responses, your pain levels have increased on days after activity. Your clinician will want to discuss this.
What this means for you
- Your body may need more recovery time between physical activities — your clinician can help adjust your routine
- The pattern is consistent enough to discuss whether your current pain management is working as well as it could
Questions you might want to ask:
Tap to add to your appointment notes
This summary was generated by AI. Your clinician will review it. You can ask for a human explanation at any time.
Key Design Principles:
- AI conclusions are always explained, never just asserted
- Patient language and clinical language are treated as equally valid inputs
- Every AI recommendation includes a “why this was flagged” disclosure
- Patients can request human review of any AI-generated finding (EU AI Act Article 22)
- Clinicians can override, annotate, and correct AI outputs with full audit logging
- Appointment prep is shared — both parties enter the consultation with the same information
These are simplified previews. Full interactive prototypes available in our complete package.
Research Interview Framework
A bidirectional research approach designed to capture both patient comprehension and clinician confidence in AI-mediated health consultations.
Bidirectional explainability requires research with both sides of the consultation:
Patients with Chronic Conditions
- Patients using digital monitoring (cardiology, orthopaedics, oncology)
- Mix of health literacy levels and digital confidence
- Include patients who have received AI-generated health summaries
- Recruit 15-20 participants across at least 2 clinical areas
Clinicians Using PROMs Platforms
- GPs, consultants, and nurse specialists reviewing AI outputs
- Mix of AI-positive and AI-sceptical clinicians
- Those using Patient-Reported Outcome Measures (PROMs) in practice
- Recruit 10-12 across primary and secondary care
Healthcare IT Leads
- Hospital teams implementing AI diagnostics or monitoring
- Compliance officers familiar with EU AI Act and GDPR health obligations
- Chief Clinical Information Officers (CCIOs)
- Recruit 5-8 from NHS trusts or equivalent institutions
Ready to make health AI legible for everyone in the room?
Our complete package includes bidirectional explainability frameworks, EU AI Act Article 13 compliance templates, and patient communication design patterns for health AI platforms.