Anchoring Enterprise GenAI with Knowledge Graphs Jonathan Lowe (Pfizer), Stephen Chin (Neo4j)
AI Summary
Video Summary: Utilization of Generative AI in Life Sciences
- Introduction:
- Speaker reflects on their roots and the importance of leadership in technology adoption, particularly generative AI (Gen AI).
- Gartner’s prediction: 30% of generative AI projects to be abandoned by 2025 due to various challenges.
- Challenges in Generative AI Projects:
- Performance issues with projects; many are not reaching production.
- Need for leadership to drive successful Gen AI initiatives.
- Importance of clear business use cases that solve real problems and are monetizable.
- Case Study:
- Jonathan describes his experience in implementing generative capabilities at a Life Sciences company.
- Focus on scaling pharmaceutical production from lab-scale to industrial scale.
- Importance of Knowledge Transfer:
- Aging workforce in manufacturing leading to loss of knowledge.
- Generative AI can help capture and transfer expertise to new employees.
- Implementation Strategy:
- Loaded millions of documents into a graph database for better management and retrieval.
- Structured document chunks to enhance search accuracy.
- Navigating Corporate Structure:
- Importance of understanding organizational hierarchy and communication challenges.
- Engagement strategies include connecting with executives and understanding their goals.
- Technology Choices:
- Emphasis on the use of graph databases for enhanced data understanding and search efficiency.
- Processing power and enhanced context retrieval through integration of graph and vector databases (“graph rag”).
- Conclusion:
- Optimization of generative AI applications can lead to significant advancements in life sciences.
- Challenges can be navigated with strategy, leadership, and the right technology.