Regulated AI & Compliance

An LLM-Assisted QMS That Auto-Scores Regulatory Readiness

Applied GenAI for compliance automation across organisations

A quality system that tells an organisation how ready it is for an audit, criterion by criterion, and uses an LLM to help close the gaps.

The problem

Checking documentation against regulations like 21 CFR Part 11 and ISO 13485 is slow, manual and easy to get wrong across many organisations.

The approach

Documentation is scored against individual criteria, such as 21 CFR Part 11 §11.10(a) to (h), and rolled up into a readiness score across Part 11, ICH E6(R3), 21 CFR Part 312 and ISO 13485. A RAG-based assistant maps evidence to criteria and drafts missing SOPs, forms and validation artifacts.

The outcome

Automated, explainable regulatory-readiness scoring with AI-drafted documents to fill the gaps.

In detail

Built a quality-management system that scores an organisation's documentation against individual regulatory criteria, for example 21 CFR Part 11 §11.10(a) to (h), and rolls them up into an overall readiness score. It covers multiple standards (21 CFR Part 11, ICH E6(R3), 21 CFR Part 312 and ISO 13485), and a RAG-based LLM assistant maps existing evidence to each criterion and drafts the missing SOPs, forms and validation artifacts.

Built with

• LLMs
• RAG
• Compliance Automation
• 21 CFR Part 11
• ICH E6(R3)
• ISO 13485

More in Regulated AI & Compliance

• NextPhase: Post-Market Surveillance for ML Models
• A 21 CFR Part 11 and ISO 13485 Compliant Electronic Document Management System

Next project: AI Receptionist for Car Dealerships

By Kanishka Ranaweera, PhD, Machine Learning Engineer in Melbourne, Victoria, Australia. hello@kanishkabuilt.me