NSF was looking to reduce both the time and risk involved in auditing new medications for safety, efficacy, and regulatory compliance. Each audit could involve tens of thousands of documents that needed to be collected, organized, version-controlled, reviewed, and summarized against strict, country-specific and international regulations. This work was highly manual, repetitive, and time-consuming, which increased the risk of human error and slowed down the overall review process. The organization wanted to: - Cut audit turnaround times so that potentially life-saving or life-improving treatments could reach patients and hospitals faster. - Free scientists and regulatory experts from low-value manual tasks so they could focus on higher-level work, such as developing regulatory strategies. - Create a scalable, repeatable way to handle growing volumes of audits without simply adding more headcount. After trying to build an AI agentic tool internally, NSF partnered with the Microsoft Cloud Accelerate Factory. Although NSF had planned for a year-long development effort, Microsoft experts helped deliver a working proof of concept in about 12 weeks, giving NSF a practical way to streamline its life sciences audits and set a foundation for broader AI use across the organization.

NSF’s Azure AI-based solution is designed to handle the heavy lifting across the audit lifecycle, from document intake to summary creation, while keeping experts in control of final decisions. Here’s how it fits into the process: 1. **Document collection and organization** - Audit materials—often tens of thousands of documents—are stored in Azure Blob Storage. - Azure Document Intelligence scans and verifies that all required documents are present. 2. **Structuring and classification** - Azure OpenAI models, working with Azure Model Context Protocol (MCP) tools, sort documents and data into the correct folders and structures that align with international regulatory frameworks. 3. **Version control and tracking** - A combination of Azure Python SDK, Azure Document Intelligence, and Azure Cosmos DB automates version tracking for each document, so NSF can maintain a clear audit trail and comply with regulatory expectations. 4. **Summarization and synthesis** - Azure Document Intelligence extracts raw text from documents. - Azure OpenAI models generate draft summaries that synthesize findings from across the entire document set. - NSF experts then review and refine these summaries, focusing on judgment and strategy rather than manual compilation. In practice, this setup has: - Cut average audit time from **4–6 weeks down to about 2 weeks**—a reduction of half or more. - Delivered what NSF describes as **near-100% truth value** in the AI-generated summaries, with humans mainly making cosmetic or style edits. The result is a reimagined audit workflow where AI manages the volume and complexity of documentation, and NSF’s specialists concentrate on higher-value regulatory and scientific work.

NSF’s Azure AI solution has delivered both operational gains and stronger controls around quality and security. **Key benefits:** - **Faster audits:** Average audit duration has dropped from **4–6 weeks to about 2 weeks**, helping clients bring medications and therapies to market more quickly. - **Reduced manual workload:** Scientists and regulatory experts spend less time on document handling and more time on higher-value activities like regulatory strategy and complex analysis. - **Improved consistency and accuracy:** The AI-generated summaries have shown **near-100% truth value**, with human reviewers mainly adjusting tone and formatting rather than correcting facts. - **Scalability across the business:** The current life sciences solution can be replicated and tailored for other areas such as medical devices, dietary supplements, and water safety, using similar structured data and regulatory patterns. - **Upskilled internal teams:** By working alongside Microsoft’s Cloud Accelerate Factory team, NSF technologists gained hands-on experience with Azure AI, prompt design, and agentic architectures, which is shaping NSF’s broader AI strategy. **Security and compliance safeguards:** - **Private, controlled environment:** Data is stored in a private tenant in SharePoint and then moved into Azure Blob Storage. All audit workflows run entirely within the Azure Cloud, with private connections to reduce exposure. - **Role-based access:** NSF uses Microsoft Entra ID and Azure role-based access control to ensure that only authorized users can access specific data and tools. - **Configurable isolation and connectivity:** Azure Model Context Protocol (MCP) Servers help manage how language models interact with tools and data, allowing NSF to decide how networked or isolated each AI solution should be. - **Aligned with Microsoft ecosystem:** As a Microsoft-focused organization, NSF benefits from the integrated security and governance features across Microsoft 365, Azure, and tools like Microsoft 365 Copilot. Together, these measures allow NSF to reshape its audit operations—speeding up reviews and reducing manual effort—while maintaining the level of security, privacy, and regulatory rigor expected in highly regulated life sciences environments.