Generative AI transforms organizations only when treated as a system, not a set of tools. This session introduces AI translation as the discipline that moves life-science enterprises from fragmented pilots to integrated value. Drawing on global projects, it examines how leadership vision, data infrastructure, and organizational learning harmonize to build capacity, steward responsible adoption, and translate models into outcomes that matter.

This presentation introduces a single AI-powered content-generation platform that integrates directly into standard word processing and spreadsheet tools. Operationalized across R&D functions, including clinical data management and medical writing, it accelerates study startup by up to four weeks. Unlike typical document-centric AI tools, the platform extends to clinical data artifacts, demonstrating how emerging agentic AI technologies enable low-barrier, multi-domain adoption of generative AI across diverse research and development workflows.

Sanofi's Global Regulatory Affairs has developed two innovative GenAI platforms to transform Common Technical Document (CTD) authoring processes. The CTD Self-Service Platform facilitates draft content generation across multiple document sections, while HAQA leverages historical health authority Q&A data to generate consistent responses. These platforms, integrated into regulatory workflows, demonstrate significant efficiency gains with 80% CTD coverage, marking a paradigm shift in regulatory documentation management.

AI now appears on nearly every product roadmap, yet turning AI capabilities into products that are trusted, governed, and actually used in practice remains a central challenge—especially in complex, regulated domains. This session takes a workflow-centric perspective, treating AI not as a standalone feature but as an integral part of end-to-end technical workflows that span people, processes, and systems. Drawing on applied experiences in digital health, governance, and AI, the talk will outline concrete design principles and patterns for integrating AI into real-world work. We will examine how to map existing workflows and identify appropriate points for AI intervention; define human–AI collaboration modes and decision rights; and redesign workflows so that AI reduces friction rather than adding steps or risk. A particular emphasis will be placed on governance: how to embed accountability, data stewardship, and feedback mechanisms into AI-powered workflows so they remain safe, auditable, and improvable over time. This session is ideal for product leaders who are being asked to “add AI” and need a structured way to respond. Attendees will leave with a practical toolkit—anchored in governance and  workflow design—that they can apply to build AI-enabled products that are reliable, responsible, and aligned with real-world practice

Outsourcing R&D to Contract Research Organizations requires precise coordination. Data, metadata, and material transfers between internal teams and CROs often create miscommunication, fragmented workflows, and extended timelines. We present an agentic AI solution that validates data return formats and empowers CRO personnel to align their outputs with internal data schemas before reaching scientists, reducing manual curation and accelerating downstream analyses.

Pharma Research-as-a-Service (RaaS) redefines drug discovery and development by uniting AI-driven automation, multi-cloud scalability, and compliance-by-design frameworks. Delivered as a secure, AWS-native SaaS platform, RaaS enables pharmaceutical and life sciences organizations to rapidly create, manage, and scale reproducible research environments across hyperscalers and specialized clouds—forming a true “science supercloud.” It accelerates innovation, enhances collaboration, reduces infrastructure overhead, and ensures traceability and regulatory alignment from concept to clinical insight.

This presentation showcases novel generative AI and interactive design platforms developed at Merck & Co Inc., for de novo design of therapeutic proteins, antibody optimization, and enzyme engineering for biocatalysis. Featuring proprietary protein language models, scalable, bespoke pipelines, and user-friendly web applications, this talk demonstrates how GenAI is being operationalized to accelerate molecular design and democratize innovation across drug discovery, vaccine development, and manufacturing workflows.

Aristotle is a next-generation AI co-scientist designed to improve the reliability and trustworthiness of AI in scientific research. The platform integrates multiple capabilities: X1 Verify evaluates experimental designs, hypotheses, and interpretations against competing paradigms; X1 Spark generates hypotheses and detailed experimental plans; X1 Search combines a language model with a scientific knowledge graph to explore and summarize literature with full reference transparency; and X1 Instant delivers rapid, accurate responses to scientific queries. Autopoiesis also introduces proprietary validation tools, including Dynamic In-line Definitions (DIDs) to clarify scientific concepts in context, and Double Check, which extracts and independently validates claims from AI outputs to reduce hallucinations and identify disputed science. Together, these capabilities provide a framework for more rigorous, transparent, and dependable AI-assisted scientific discovery.

As GenAI moves deeper into clinical, regulatory, and R&D workflows, teams face major validation barriers due to PHI exposure risks and proprietary data restrictions. AI-generated synthetic data offers a safe, scalable alternative—producing high-fidelity structured and unstructured datasets with referential integrity. This talk will demonstrate how synthetic data closes validation and compliance gaps, de-risks deployment, and enables reliable, audit-ready GenAI systems across scientific, clinical, and operational environments.