Inspiration

As a Senior Clinical Evaluation Medical Writer with over 20 years in pharmacy and hands-on experience authoring CER, PMCF, and PSUR documents for oncology and Class III medical devices, I repeatedly saw the same problem: small and mid-sized medical device companies often lack large in-house regulatory teams, yet face the same compliance scrutiny as major manufacturers. Reviewing these documents manually is slow, expensive, and error-prone — even small gaps can delay approval or trigger regulatory findings. I wanted to combine my real-world regulatory expertise with AI to make expert-level compliance review faster and more accessible.

What it does

MedAd Compliance AI reviews clinical and regulatory documents — Clinical Evaluation Reports, Post-Market Clinical Follow-up reports, and Periodic Safety Update Reports — and flags compliance gaps, inconsistencies, and missing elements against EU MDR and ICH-GCP standards. Instead of a generic AI writing tool, it's grounded in actual clinical evaluation practice, helping QA/RA teams catch issues before submission rather than after rejection.

How we built it

I built MedAd Compliance AI using Lovable, integrating Gemini to power the document analysis engine. My role was to define the regulatory logic, compliance criteria, and review structure based on real EU MDR and ICH-GCP requirements, while Gemini handles the natural language understanding and document parsing.

Challenges we ran into

Translating years of hands-on regulatory writing experience into clear, structured logic an AI model could apply consistently was the biggest challenge — regulatory compliance isn't always black-and-white, and capturing that nuance took careful iteration.

Accomplishments that we're proud of Building a working, publicly usable tool — not just a concept — that reflects genuine industry expertise rather than generic AI output.

What we learned That the real value of AI in specialized fields comes from pairing it with deep domain expertise, not replacing it.

What's next for MedAd Compliance AI Onboarding real medical device clients, gathering feedback to refine accuracy, and expanding coverage to additional regulatory document types.

Built With

  • ai
  • compliance
  • gemini
  • healthcare
  • lovable
  • medical
  • regulatory
Share this project:

Updates

posted an update

Submission complete!

MedAd Compliance AI is now submitted for the Build with Gemini XPRIZE. The tool uses Google's Gemini API to help small medical device manufacturers catch non-compliant marketing claims before publication—flagging issues against EU MDR requirements and generating compliant rewrites in real time.

Built as an extension of an existing portfolio project, with new features developed and manufacturers actively contacted during this hackathon period.

Try it: ileana-mazilu.lovable.app/medad-ai

Log in or sign up for Devpost to join the conversation.