A core group of Boston’s forward-thinking senior data executives gathered at CDO Magazine’s executive boardroom dinner October 24 to network, peer mentor, and share their perspectives on “GenAI: How You Are Exploring or Using It.”

Data leaders from various sectors, including government, pharmaceutical, financial services, and biotech discussed topics ranging from how to create business value and prioritize investments to how to manage expectations, address governance, and upskill human resources, moderated by Matt Sweetnam, AHEAD Chief Architect.

Leaders on CDO Magazine’s Global Editorial Board shared several key points and takeaways. Krishna Valluru, Fidelity Investments VP of Advanced Data Analytics, sums up what his peers are facing: “It’s become apparent that they are encountering significant challenges when it comes to utilizing data for GenAI use cases.”

While actively considering ethical implications within their models, he says the most prominent obstacles stem from data quality, trust, and skill, with a focus on data preprocessing. Valluru adds that many use cases they face revolve around content generation for marketing, FDA regulatory filings, and summarization.

Sravan Kasarla, Thrivent Chief Data Officer and Head of AI/ML, warns practitioners to be careful with GenAI, saying, “Let’s not make it a hammer you’re going to use now and everything becomes a nail. Make sure that you’re applying it for the right use case.” He elaborates the importance of:

1. Managing expectations of the organization about the early stages of Generative AI. It is important to make sure that generative AI is understood for what it is and what it is not.
2. Creating an environment where you’re able to innovate but have guardrails to ensure that it is getting applied to the right use cases. It needs to be evaluated to ensure you are gaining value because there is cost.
3. Avoiding betting on a single LLM or foundational model or method since technology is changing so rapidly. Even the biggest names like OpenAI, Google Bard, AWS Bedrock, and LLaMA are fundamentally going to continue to change.

Specifically, Kasarla recommends, “You create a layer of abstraction where you can use the power of the large language model, whether it is the best NLP or just search and summarization. But, design it so that you can abstract a way to a different, more powerful, or better model.”

He adds, “The use cases which are the most successful, in my opinion, are leveraging the NLP power of the large language model for search and summarization, and text creation. Know your data that you are passing through a large language model to have a better context and results you can stand behind.”

Mamta Singh, Commonwealth of Massachusetts Deputy Chief Data Officer, suggests, “While there’s no denying the potential hitches in terms of ethics and legality, the establishment of a GenAI (generative AI)  governance committee can address these concerns. A significant boon of GenAI lies in its ability to eliminate repetitive tasks, thus liberating human resources for more meaningful endeavors.

“While the journey of integrating GenAI into our lives will come with its set of challenges, with the proper measures and intentions, it promises a brighter, more efficient future,” Singh adds.

According to Barbara Latulippe, Takeda Head of Enterprise Data, key takeaways were:

1. The key to starting your GenAI journey is a central registration of use cases that can be prioritized and scaled across multiple business units. Emerging themes are content generation and knowledge extraction which significantly improve productivity and value realization as a co-pilot with a human in the loop.
2. Creating a GenAI community of practice that taps into the wisdom of global stakeholders, consumers, and producers helps surface opportunities that scale, elevate enterprise capabilities, and foster a culture of collaboration.
3. Developing a risk framework on the ethical, explainable, and responsible use of AI/GenAI in addition to a strong foundation of data quality should be the first priority to achieve any AI ambition.”

Data executives attending the CDO Magazine Boston Executive Boardroom Dinner included:

Sravan Kasarla, Chief Data Officer and Head of AI/ML; Barbara Latulippe, Takeda Head of Enterprise Data; Jennifer McGhee, MFS Investment Management Vice President, Sr. Director of Data Strategy & Governance; Cliona Molony, IDEXX Chief Data Officer, VP R&D; Raj Nimmagadda, Sanofi Chief Data Officer R&D, Data and Data Sciences; Mamta Singh, Commonwealth of Massachusetts Deputy Chief Data Officer; Jane Urban, Otsuka Pharmaceutical Companies Vice President, Customer Engagement Operations; Krishna Valluru, Fidelity Investments VP, Advanced Data Analytics; Dave McEachern, AHEAD Managing Director; Matt Sweetnam, AHEAD Managing Director; Ben Prescott, AHEAD Principal Technical Consultant – Data Science; SteveWanamaker, CDO Magazine Founder & Publisher; and Camille Prado, CDO Magazine Global Editor.