Pharmaceutical companies cut CTD authoring time 60–65% with Accenture's AWS generative AI solution

Pharmaceutical companies face an extraordinarily labor-intensive process of manually creating Common Technical Documents (CTDs) for FDA submissions, requiring up to 100,000 hours per year for a typical large pharma company. The CTD contains over 100 highly detailed technical reports compiled during drug research and testing, making the process both time-consuming and prone to errors.

Accenture built a generative AI-based regulatory document authoring solution on AWS that automatically extracts key data from testing reports and generates CTDs in the required format using Amazon SageMaker JumpStart with AI21 Jurassic Jumbo Instruct and AI21 Summarize models. The system is built on AWS Well-Architected principles with encryption and auditability controls, and allows users to review and edit computer-generated documents before submission.

Early testing demonstrated a 60–65% reduction in the time required for authoring CTDs, exceeding preliminary estimates of 40–45% reduction. The solution accelerates new drug approvals by compressing the document compilation process while maintaining compliance with FDA format requirements.

• Generative AI can eliminate the majority of manual effort in regulatory document creation, potentially saving tens of thousands of hours annually per large pharma company
• A human-in-the-loop review layer is essential for regulated industries, even with high automation fidelity
• AWS Well-Architected principles (security, auditability, encryption) are table-stakes for handling sensitive pharma submission data