Short answer: European healthcare companies can deploy AI in mobile apps fully GDPR-compliant by running inference on-device — data minimization is satisfied structurally because personal data never transmits outside the device. Fixed price, 4–6 weeks, money back.
Your DPO's legal opinion is that processing patient health data through a US cloud LLM requires an adequacy decision that doesn't currently exist. Your clinicians need the AI feature anyway.
The legal opinion is correct. The clinical need is also real. These two facts together define the architecture you need to build - not a workaround for the legal opinion, but a solution that makes the legal opinion irrelevant.
What decisions determine whether this project ships in 6 weeks or 18 months?
Four decisions determine whether this project navigates the GDPR special category framework or stalls in a legal loop that outlasts your product roadmap.
Special category processing basis. Article 9 requires an explicit lawful basis for processing health data with AI. Legitimate interest doesn't cover health data. Your legal team needs to determine whether the "provision of health care" exemption applies to your use case, or whether explicit patient consent is required for each AI processing activity. This decision has to be made before the model is configured, not after the first compliance review.
On-device vs EU-hosted. On-device processing eliminates the transfer question entirely. Patient data that never leaves the device has no transfer mechanism to document. An EU-hosted model on a compliant cloud provider requires a DPA, a valid transfer mechanism, and a sub-processor audit trail. Your compliance team's risk tolerance - and your DPO's past decisions on similar questions - determines which path gets to production faster.
Data subject rights for automated processing. GDPR gives patients the right to understand automated decisions that affect them. An on-device model that assists with clinical documentation still needs a mechanism for the patient to understand what the model processed and what it produced. The disclosure and rights architecture has to be designed into the app, not appended as a privacy policy update post-launch.
Pseudonymization before inference. Some healthcare AI tasks - summarizing clinical literature, classifying symptoms against a reference set, suggesting ICD codes from a symptom description - can run on pseudonymized data without losing clinical utility. If your data science team confirms this is true for your specific task, you may not need on-device at all. You need a pseudonymization step before any inference call. Your legal team and your DPO need to confirm the pseudonymization standard before this architecture is chosen.
Most teams spend 4-6 months discovering these decisions by building the wrong version first. A team that has shipped this before compresses that to 1 week.
On-Device AI vs. Cloud AI: What's the Real Difference?
| Factor | On-Device AI | Cloud AI |
|---|---|---|
| Data transmission | None — data never leaves the device | All inputs sent to external server |
| Compliance | No BAA/DPA required for inference step | Requires BAA (HIPAA) or DPA (GDPR) |
| Latency | Under 100ms on Neural Engine | 300ms–2s (network + server queue) |
| Cost at scale | Fixed — one-time integration | Variable — $0.001–$0.01 per query |
| Offline capability | Full functionality, no connectivity needed | Requires active internet connection |
| Model size | 1B–7B parameters (quantized) | Unlimited (GPT-4, Claude 3, etc.) |
| Data sovereignty | Device-local, no cross-border transfer | Depends on server region and DPA chain |
The right choice depends on your compliance constraints, query volume, and task complexity. Wednesday scopes this in the first week — before any code is written.
Why is Wednesday the right team for on-device AI?
We built Off Grid because we hit every one of these problems in production. Off Grid is the fastest-growing on-device AI application in the world, with 50,000+ users running it today.
It's open source, with 1,650+ stars on GitHub and contributors from across the world. It has been cited in peer-reviewed clinical research on offline mobile edge AI.
Every decision named above - model choice, platform, server boundary, compliance posture - we have made before, at scale, for real deployments.
How long does the integration take, and what does it cost?
The engagement is four sprints. Each sprint is fixed-price. Each sprint has a named deliverable your team can put on a roadmap.
Discovery (Week 1, $5K): We resolve the four decisions - model, platform, server boundary, compliance posture. Deliverable: a 1-page architecture doc your CTO can take to the board and your Privacy Officer can take to Legal.
Integration (Weeks 2-3, $5K-$10K): We ship the on-device model into your app behind a feature flag. Deliverable: a working build your QA team can test against real workflows.
Optimization (Weeks 4-5, $5K-$10K): We hit the performance and compliance targets from the discovery doc. Deliverable: benchmarks signed off by your team.
Production hardening (Week 6, $5K): Edge cases, OS version coverage, app store and compliance review readiness. Deliverable: shippable build.
4-6 weeks total. $20K-$30K total.
Money back if we don't hit the benchmarks. We have not had to refund.
"Retention improved from 42% to 76% at 3 months. AI recommendations rated 'highly relevant' by 87% of users." - Jackson Reed, Owner, Vita Sync Health
Is on-device AI right for your organization?
Worth 30 minutes? We'll walk you through what your version of the four decisions looks like, what a realistic scope and timeline would be for your app, and what your compliance posture and on-device target mean in practice.
You'll leave with enough to run a planning meeting next week. No pitch deck.
If we're not the right team, we'll tell you who is.
Book a call with the Wednesday team
Frequently Asked Questions
Q: How does on-device AI satisfy GDPR data minimization for healthcare apps?
GDPR Article 5(1)(c) requires that personal data be limited to what is necessary. An on-device model that processes data locally and produces an inference without transmitting the input satisfies minimization structurally — no personal data reaches any third-party processor.
Q: Does running AI on a European cloud server satisfy GDPR, or is on-device required?
EU-hosted cloud AI satisfies the cross-border transfer requirement but still requires a DPA with the provider. On-device eliminates the DPA requirement for the inference step entirely. For DPOs who have already rejected cloud AI configurations, on-device is typically the faster path to sign-off.
Q: How long does GDPR-compliant on-device AI take for a healthcare app?
4–6 weeks. Week 1 resolves: data minimization architecture, model provenance, cross-border transfer elimination, and customer disclosure language. Disclosure drafting runs in parallel with the technical build — legal review doesn't become the long pole.
Q: What does GDPR-compliant on-device AI cost?
$20K–$30K across four fixed-price sprints, money back if benchmarks aren't met.
Q: What model provenance documentation do DPOs require for on-device AI?
DPOs ask whether the model was trained on personal data subject to GDPR. Open-source models with published data cards (Mistral, Gemma, Phi, LLaMA) are defensible. Closed commercial models with opaque training data require a much longer legal analysis. Wednesday defaults to open-source with published training documentation for all GDPR deployments.
Top comments (0)