Getty Images/iStockphoto

Elsevier launches PharmaPendium AI to boost R&D, compliance

Elsevier debuts PharmaPendium AI, a GenAI tool that helps regulatory, preclinical and clinical teams access FDA, EMA and other regulatory data to accelerate R&D and compliance.

Today, Elsevier announced the launch of PharmaPendium AI, a generative AI assistant designed to enhance the way regulatory specialists and R&D teams access and analyze data. Built upon the company's existing PharmaPendium platform, this AI-driven tool generates citation-backed answers to complex FDA and European Medicines Agency queries (EMA) within seconds.

"PharmaPendium AI transforms how regulatory affairs professionals, as well as preclinical and clinical teams, access and analyze data from FDA and EMA regulatory documents and numerous additional sources," Olivier Barberan, Director of Translational Medicine Solutions at Elsevier, wrote in an emailed statement. "These benefits apply across all drug modalities, including small molecules, monoclonal antibodies, antibody–drug conjugates and cell and gene therapies, and across all therapeutic areas."

The system uses retrieval-augmented generation (RAG) and natural language processing to help users navigate regulatory complexity with plain-language queries.

"Pharma teams are incorporating PharmaPendium AI into their R&D and regulatory processes by embedding it directly into existing workflows," Barberan explained. "Within large organizations, regulatory teams often need to sift through millions of pages of documentation from the FDA, EMA and other agencies. PharmaPendium AI enables them to navigate this complexity with natural language search, providing intuitive and rapid access to regulatory-grade information with traceable references."

Early access testing showed efficiency gains of up to 66% in search and review sessions, but Barberan noted that the broader value lies in accuracy.

"While saving time is valuable, the true game-changer is PharmaPendium AI's ability to ensure that critical information is not missed," Barberan said. "This reduces the risk of oversight, enhances decision quality and ultimately shortens the regulatory cycle, leading to faster approvals and significant cost savings."

PharmaPendium AI is being applied at multiple points across the development and regulatory lifecycle, including preclinical development and Investigational New Drug planning, clinical trial design and safety profiling, regulatory submissions and lifecycle management.

According to Barberan, the tool has been tested by a wide range of organizations, "from large multinationals to mid-sized and smaller companies," and across different modalities, including biologics and combination therapies.

To maintain reliability, PharmaPendium AI generates responses exclusively from PharmaPendium's content, comprising over 5 million pages of FDA approval packages, EMA filings, Advisory Committee transcripts and Meyler's Side Effects of Drugs 

"With our extended RAG approach, we deliberately exclude prior knowledge embedded within the large language model to ensure the integrity and accuracy of the information," Barberan reiterated. "By relying solely on the knowledge contained within the regulatory documents, we effectively minimize the risk of hallucinations and misinformation."

Human oversight provides an additional layer of quality control, and results can be delivered in multiple formats, from concise summaries to submission-ready tables aligned with regulatory standards.

Barberan added that the platform was developed in collaboration with regulatory and R&D professionals across the pharma industry, in line with Elsevier's Responsible AI and Privacy Principles, which emphasize transparency, data security and confidentiality.  

"All user interactions are private, with no data used to train external models," he highlighted.

Beyond privacy and compliance safeguards, PharmaPendium AI's ability to compare new therapies with existing drugs offers companies a way to highlight differentiation opportunities while reducing development risks. By incorporating insights from past FDA and EMA decisions, this AI-driven technology has the potential to strengthen approval strategies and post-marketing planning, Barberan concluded.

Alivia Kaylor is a scientist and the senior site editor of Pharma Life Sciences.

Next Steps

FDA launches gen AI Elsa to support clinical, regulatory tasks 

Dig Deeper on Pharmaceuticals