Intelligent Authoring Using GenAI and Structured Content

Business Challenges

The biopharma industry operates in a highly regulated environment with constantly evolving guidelines. To navigate this landscape, leveraging advanced technologies is essential for streamlining content creation and ensuring compliance. To achieve transformative operational efficiencies, Generative AI (GenAI) is proving to be a key tool, revolutionizing content generation, regulatory information management, and submission processes.

The ArborSys Solution

ArborSys worked closely with our client to implement a pilot POC of InteliNotion. Building on top of the core structured content management foundation, InteliNotion has implemented GenAI capabilities to automate the draft document generation process, and support AI assisted authoring within Microsoft Word.

Our client has previously implemented InteliNotion structured content authoring for several clinical documents, such as the protocol, clinical study report (CSR) body and Statistical Analysis Plan (SAP). We worked with our client to deploy the GenAI version of the InteliNotion platform to extended the configurations to incorporate GenAI agents for client preferred Large Language Models (LLMs), document template level GenAI prompts for the draft document generation, and the best practices for AI assisted authoring.

ArborSys supported our client’s project team by implementing InteliNotion for an initial pilot and training the client medical writing and project team with the best practices for prompt engineering, prompt configurations and incorporating AI- assisted authoring into their existing authoring and review processes. The Protocol, Statistical Analysis Plan (SAP), and Clinical Study Report (CSR) had been previously structured and configured with appropriate reuse policies, and libraries were established for standardized content that could be applied across these documents. The existing configurations were extended to support GenAI, configure and test with applicable LLMs.

Business Value

By shifting from standard structured content authoring processes to an GenAI and AI-assisted authoring approach, medical writers expect to improve and streamline the generation, review, and document finalization process up to 45% from current timelines. This transformation is expected to lead to increased efficiency and consistency, reducing the risk of errors and enhancing traceability. The combined solution of structured content and GenAI facilitates content reuse and traceability of the AI generated content.

Additionally, structured content offers a robust framework for configuring component-level prompts, ensuring segmentation, detailed context, precision, consistency, and compliance. By breaking information into distinct components, structured content enables the creation of detailed prompts for each segment.

Component-level prompting enhances GenAI accuracy by focusing on specific components, providing context-specific information, incorporating metadata, specifying data requirements, and enabling customization.

InteliNotion’s capabilities to configure precise and contextual component-level prompts, as well as the configurability of multiple LLMs fit for purpose (for example, different LLMs for generating text or text from images/ tables) provides a highly flexible way to configure the right models with contextual prompts to generate higher quality and accurate draft content. Models can also be adopted and configured easily as they mature.