Credence

Pistoia Alliance Expands AI Initiatives in Drug Development and Pharmacovigilance


Key Points

  • Pistoia Alliance launched AIPAKA, an AI knowledge assistant, and new collaborative programs at its largest Boston conference to date
  • ClinSilico consortium will focus on AI in silico clinical development, including digital twins and adaptive trial design
  • New AI in Pharmacovigilance community and FAIR Business Value Framework address data quality and return on investment for AI adoption

The Pistoia Alliance has introduced several collaborative industry programs focused on artificial intelligence (AI) in clinical development, pharmacovigilance, FAIR data practices and laboratory digital transformation. The announcements were made during the organization’s largest annual Boston member conference to date, which brought together pharmaceutical, biotechnology and technology representatives from companies including AbbVie, Bayer, Eli Lilly, GSK, Johnson & Johnson, Pfizer, Roche, Takeda and UCB.

The nonprofit life sciences organization also introduced AIPAKA, an AI-powered knowledge assistant designed to help members access more than a decade of industry research, project findings and shared expertise. The tool allows users to submit questions in natural language and retrieve information from Alliance webinars, best-practice resources, technical guides and previous collaborative initiatives. The new platform is intended to make the organization’s collective knowledge more accessible, including resources such as its FAIR Toolkit and a newly developed Project and Community Roadmap that tracks ongoing activities and milestones.

Dr. Becky Upton, president of the Pistoia Alliance, said discussions at the conference highlighted the importance of collaboration between scientific, data and technology teams as pharmaceutical companies expand their use of AI. She noted that the limitations of AI adoption in research and development increasingly involve data quality, supporting infrastructure and organizational practices rather than the capabilities of the models themselves. Addressing those challenges requires industry-wide cooperation and common approaches to data management and technology implementation.

Medicinal chemist and industry commentator Derek Lowe reinforced this perspective during his closing remarks on the first day of the conference. He emphasized that while AI can help identify patterns and extract insights from existing information, human expertise remains essential for interpreting results, applying scientific judgment and determining where the technology can deliver meaningful value.

Among the initiatives announced or advanced at the Boston meeting was ClinSilico, a consortium being developed by the Pistoia Alliance in partnership with Cambridge University Health Partners. The program will focus on AI applications in in silico clinical development, including digital twins, adaptive clinical trial design and AI agents. Participating organizations include the Cambridge Centre for AI in Medicine, the MRC Biostatistics Unit and UK CEiRSI. The Alliance will contribute its collaborative infrastructure and precompetitive partnership model. ClinSilico is part of the organization’s strategic effort to accelerate late-stage research and development.

The Alliance also announced progress on its AI in Pharmacovigilance Community of Experts, which will provide a forum for pharmaceutical companies to exchange experience and develop best practices for implementing AI in drug safety operations. Roche, CSL Behring and Ergomed are sponsoring the initiative, with AstraZeneca participating as a contributor. The Alliance is seeking additional sponsors and industry participants to expand the program.

Data standardization and accessibility remain central to the Alliance’s work, particularly as pharmaceutical companies seek to establish the data foundations required for broader AI adoption. Its FAIR for Pharma Community of Experts has developed a FAIR Business Value Framework intended to help organizations assess the financial and operational benefits of making data findable, accessible, interoperable and reusable. The framework connects 38 value parameters across four categories: effectiveness, trust, cost and speed. An accompanying calculator allows members to estimate the return on investment associated with FAIR initiatives and develop business cases for investments in data infrastructure. The framework is available at no cost, while the calculator is accessible to Alliance members. The organization is encouraging companies to share their experiences using the resources to support further development and adoption.

Further information is available at www.pistoiaalliance.org.