MERCK & THIGA

From Raw Data to Strategic Insight: Turning AI into a Competitive Advantage at Merck

Product Craft
Articifial Intelligence & Data
Healthcare & Pharma
merck

Context

Merck, a global German pharmaceutical company, conducts cutting-edge research to develop new treatments for chronic and often incurable diseases. Its scientific teams rely on vast and complex data sets to understand the real-world impact of these treatments on patients and to improve clinical decision-making.

In this context, Thiga supported Merck in structuring its product governance, stabilizing its generative AI solution, and scaling an analytics platform designed to transform patient and caregiver insights into a true strategic asset.

Challenges

  • Managing data complexity at scale: unify and validate multiple heterogeneous sources, from physicians and scientific literature to social media, while integrating LLM models with evolving and sometimes unstable behaviors.
  • Aligning global teams around a shared vision: balance centralized governance with regional needs to ensure coherence, compliance, and local efficiency.
  • Securing and stabilizing the generative AI solution: maintain reliability and performance in a fast-changing technological environment.
  • Building a clear and shared product approach: establish a product culture within a still-maturing environment, aligning teams around common goals and priorities.

Our approach

  • Square - Red 30

    Expanding the AI analytics product

    Developed new use cases through a Qlik dashboard and a Streamlit natural language app, coordinated with the AI platform to deliver a unified experience across products.

  • Square - Red 26

    Driving discovery and user-centered prioritization

    Conducted user interviews, mapped use cases, and collaboratively defined key features with R&D, Medical, and Market Access teams.

  • Square - Red 29

    Industrializing AI-powered workflows

    Designed automated analysis pipelines (Q&A, summarization, clustering, NER, sentiment analysis), implemented LLM chains, optimized prompting, and monitored performance through LangFuse.

  • Square - Red 14

    Ensuring proactive incident management and transparent communication

    Deployed a monitoring and alerting system to anticipate issues and maintain continuous, open communication with users.

  • Square - Red 07

    Defining a long-term vision and migration toward Agentic AI

    Contributed to Merck’s strategic roadmap for adopting an Agentic AI platform, ensuring alignment with internal standards and R&D initiatives.

Our impact

  • Accelerated time-to-insight: report generation accelerated by 4–5x and autonomous access to personalized analyses through natural language queries.
  • Enhanced clinical and strategic decision quality: faster validation or invalidation of hypotheses through quantitative data and improved integration of patient criteria aligned with regulatory expectations.
  • Strengthened reliability and product maturity: significant reduction in debugging time, improved communication with users, and the emergence of a strong product culture around AI.

Would you like to know more about this project or another?