Manik Gupta, Chief Analytics and Insights Officer for North America Consumer Health at Bayer, speaks with Tracy Ring, Chief Data Officer and Global Generative AI Lead for Life Sciences at Accenture, in a video interview about his professional background and role at Bayer, and the organization’s experience of using LLMs since much before the advent of ChatGPT.
Bayer Consumer Health division has over 170 consumer health brands that cover areas such as dermatology, nutritional supplements, pain management, cardiovascular risk prevention, digestive health, cough, cold, and allergy care.
Gupta has a mixed background in strategy, technology, and data science across verticals like healthcare, CPG, retail, financial services, telco, and advisory services. In his role at Bayer, he oversees multiple AI Communities of Practice, including human truths, consumer engagement, enterprise, business intelligence, advanced analytics, and data science. and the AI practice in Canada.
Sharing his take on the emergence of generative AI (GenAI) as a business enabler and disruptor, Gupta says that GenAI has the potential to change the anatomy of work across every industry and every function, augmenting the capabilities of individual workers by automating activities.
He however maintains that the full realization of benefits from advancing GenAI will require managing inherent risks and determining the new skills and capabilities required in the workforce. Gupta states that eventually, it will be less about the technology and more about how humans and businesses are organized around it.
Gupta reveals that Bayer started experimenting and embedding large language models (LLMs) much before ChatGPT made it popular. This has helped the company get a clear understanding of important but unmet consumer needs.
He mentions leveraging BioGPT (GenAI for biomedical text generation and mining) on about 47 million publicly available medical citations. This improved the productivity of a medical affairs team by 100x and accelerated the process of connecting the right health outcomes and the right molecules, to address unmet consumer needs.
Highlighting another example Gupta mentions using about 9 billion rows of structured data for NLP processing. This helped practitioners query the data using a search bar and understand the market dynamics quickly. While Gupta says that there are additional Gen AI use cases under development, he also suggests that the number of types of use cases is only limited by imagination.
Speaking further on the topic, he says that beyond the GenAI hype cycle and usage, there are real risks like copyright infringement and technical challenges like limited GPU supply.
In conclusion, Gupta mentions the following rules of engagement at Bayer:
Protect proprietary data and thoughtfully deploy new technologies behind firewalls
Continuously calibrate and recalibrate risk-reward
Build solutions to get the unique advantage
buy off-the-shelf solutions where it makes more sense
CDO Magazine appreciates Manik Gupta for sharing his insights and data success stories with our global community.