JNP Medi announced on the 21st that it will release a beta version of Maven Safety AI Import, a new feature that integrates AI-powered data extraction into its pharmacovigilance (PV) support solution.
JNP Medi’s existing solution, Maven Safety, has supported the full range of pharmacovigilance tasks, from drafting and reviewing adverse event reports to approval and regulatory reporting. The newly introduced feature applies AI-based data extraction technology to the existing platform to reduce the burden of repetitive data entry.
Pharmacovigilance involves collecting and analyzing adverse events associated with the use of medicines to identify and prevent emerging safety concerns. It is used to continuously monitor potential side effects after a drug enters the market and to establish appropriate safety measures when necessary.
Pharmacovigilance teams typically enter information from adverse event reports into safety databases and compare the entries with the original documents to identify omissions or errors. Non-standardized documents can be particularly time-consuming, as staff must manually review and enter the information.
Maven Safety AI Import allows users to upload adverse event reports in PDF format. The AI extracts key information and automatically populates fields in accordance with the International Council for Harmonisation of Technical Requirements for Pharmaceuticals for Human Use (ICH) E2B(R3) electronic reporting standard. Users can compare the extracted information with the original document, make any necessary corrections and generate the final dataset.
In an internal test covering 106 data fields, JNP Medi recorded an available-data entry rate of 87.8% based on the documents used and an accuracy rate of 95.3% for the values entered.
The company expects the feature to improve workflow efficiency in settings where small teams are responsible for data entry, quality control and regulatory reporting.
Currently available in beta, the feature can be used by existing Maven Safety users with data-import permissions. JNP Medi plans to incorporate user feedback, improve extraction accuracy and usability, and officially launch the feature in the first half of 2027.
“We aim to reduce the repetitive data-entry burden on pharmacovigilance professionals so they can focus on reviewing safety information and making informed decisions,” said JNP Medi CEO Jung Kwon-ho. “We will continue improving the feature’s accuracy and ease of use in response to the needs of professionals in the field.”
