By Sagar Shankaran, Founder of CallSphere
Some AI voice and chat agents are software-as-a-medical-device under EU MDR (2017/745). Here is how Rule 11 classification, IEC 62304, and the AI Act interact for healthcare voice in 2026.
Key takeaways
Most AI voice agents are not medical devices. The ones that triage symptoms, monitor adherence, or output a treatment recommendation can be — and EU MDR Rule 11 plus the AI Act now apply in tandem.
Regulation (EU) 2017/745 (MDR) governs medical devices on the EU market. Software qualifies as a medical device when it is intended for diagnosis, prevention, monitoring, prediction, prognosis, treatment, or alleviation of disease or injury. Rule 11 of MDR Annex VIII is the headline classification rule for software: it is at least Class IIa, Class IIb if it can drive decisions about diagnosis or therapy of serious conditions, and Class III where errors could cause life-threatening consequences. Class IIa, IIb, and III all require a Notified Body conformity assessment.
The technical baselines are well-established: ISO 13485 for the quality system, IEC 62304 for software lifecycle, ISO 14971 for risk management, ISO 15223-1 for symbols, IEC 82304-1 for health software safety, and IEC 62366-1 for usability. The AI Act layers on top: AI-enabled SaMD is automatically a high-risk AI system under Annex III, and from August 2, 2027 the AI Act fully applies to AI safety components of MDR/IVDR products requiring third-party conformity assessment.
Hear it before you finish reading
Talk to a live CallSphere AI voice agent in your browser — 60 seconds, no signup.
Step one is the qualification call: is the AI voice or chat doing something with medical purpose? Reminders, scheduling, payment IVR, and benefits navigation are not. Symptom triage, sepsis screening, suicide-risk assessment with clinical action, and titration coaching often are. Step two is classification: most clinically active SaMD lands at IIa or IIb under Rule 11.
Once in scope: ISO 13485 quality system, IEC 62304 software-lifecycle artifacts (development plan, requirements, architecture, unit/integration/system test, problem reports, configuration management), ISO 14971 risk file with intended-use scope, hazard analysis, and benefit/risk acceptability, clinical evaluation under MDCG 2020-1 and MDCG 2020-6, post-market clinical follow-up, and the matching AI Act Annex IV technical documentation. Notified Body audit, CE marking, EUDAMED registration, and a Person Responsible for Regulatory Compliance (PRRC) follow.
For non-SaMD use cases — appointment booking, intake, eligibility, billing — CallSphere serves as a HIPAA and SOC 2 aligned voice and chat platform with 37 agents, 90+ tools, 115+ DB tables, 6 verticals, 50+ businesses, and a 4.8/5 rating. The Healthcare Voice Agent's 14 tools include intake, scheduling, insurance check, refill nudges, and crisis routing — none of which by themselves cross into MDR scope when implemented as routing and operations rather than diagnosis or therapy. The encrypted PostgreSQL healthcare_voice database, full post-call analytics, sentiment, lead score, AI summary, and audit trail map cleanly to ISO 14971 risk evidence and IEC 62304 traceability if a customer chooses to bring SaMD use cases inside its own quality system. Behavioral-health deployments anchor at /lp/behavioral-health and explicitly route crisis events to licensed humans rather than offering clinical diagnoses. Pricing $149 / $499 / $1,499; 14-day trial; 22% affiliate. Industry hubs at /industries/healthcare and /industries/behavioral-health.
flowchart LR
A[AI Voice Use Case] --> B{Medical\nPurpose?}
B -- No --> C[Out of MDR]
B -- Yes --> D[Rule 11\nClassify]
D --> E[ISO 13485 QMS]
E --> F[IEC 62304 SLC]
F --> G[ISO 14971 Risk]
G --> H[Notified Body]
H --> I[CE + EUDAMED]
I --> J[AI Act\nAnnex IV]
Is appointment scheduling SaMD? No. Operational software without medical purpose is out of scope.
Still reading? Stop comparing — try CallSphere live.
CallSphere ships complete AI voice agents per industry — 14 tools for healthcare, 10 agents for real estate, 4 specialists for salons. See how it actually handles a call before you book a demo.
Does CE under the old MDD still count? Legacy MDD certificates expire on the published transition timeline; renew under MDR.
Can we use a generic LLM as the inference engine? Yes, but it becomes part of the medical device and inherits all obligations.
Does the AI Act replace MDR? No. They stack. Comply with both.
Written by
Sagar Shankaran· Founder, CallSphere
Sagar Shankaran is the founder of CallSphere, where he builds production AI voice and chat agents deployed across healthcare, hospitality, real estate, and home services. He writes about agentic AI, LLM engineering, and shipping voice agents that handle real calls in production.
See how AI voice agents work for your industry. Live demo available -- no signup required.
AWS HealthScribe became the open scribe layer EHR vendors built on top of in 2026. Here's the API surface, the per-encounter pricing, the BAA terms.
Apollo, Manipal, and Narayana scaled AI agents across Bangalore in 2026. Here's the deployments across radiology, intake, and follow-up, the costs.
Notable's AI agents now handle scheduling, intake, and revenue cycle for 6,000+ clinics in 2026. Here's the multi-agent architecture, the per-clinic pricing.
Abridge raised $250M in April 2026 at a $2.7B valuation. We break down the deployment numbers, the EHR integrations across Epic and Cerner. The Q2 2026 buyer briefing.
Enterprise CIO Guide perspective on Hippocratic AI's deployment numbers show healthcare voice agents are moving from pilot to production across major US health systems.
Home health agencies use AI voice agents for daily check-ins on 2.5M+ patients in 2026. Here's the deployments at major agencies, the per-patient pricing.
© 2026 CallSphere LLC. All rights reserved.