ESI Team

How to Train Your Generative AI

February 14, 2024
Share this blog post

On the 34th episode of Enterprise Software Innovators, host Evan Reiser (Abnormal Security) talks with Marty Brodbeck, CTO of Priceline. Priceline is an online travel agency that enables users to book hotels, flights, and rental cars across 400 airlines and 300,000 hotels in over 200 countries worldwide. Marty shares his thoughts on organizational frameworks to implement AI effectively, underestimated use cases for generative AI, and how Priceline's new chatbot, Penny, uses the latest generative AI capabilities to transform the customer experience.

Marty's insights into the organizational frameworks necessary for effective AI implementation are pivotal. Transitioning to AI is not just about adopting new technologies; it involves a holistic transformation of the organization's processes and mindset. Marty emphasizes the importance of a solid organizational infrastructure supporting AI's demands. He speaks to creating an environment where rapid experimentation and agile development are the norms. This approach is vital for businesses looking to harness AI's full potential, as it allows for continuous learning and adaptation, which is crucial in the ever-evolving landscape of AI technologies. "These large language models don't work right out of the box. There's a corpus of information that they get trained on that isn't necessarily accurate, so you have to complement the capabilities you're going to get from a large language model with your data and eliminate the amount of hallucinations that can happen when you marry the two together."

As an AI enthusiast, Marty also discusses the underestimated potential of generative AI capabilities in business areas that generally don't dominate the headlines. While most discussions focus on the flashy applications, he points to more practical yet often overlooked uses such as customer service automation and legacy system modernization. In customer service, for instance, AI can analyze vast amounts of historical data to provide more efficient and accurate responses, significantly reducing operational costs. Similarly, generative AI can play a critical role in translating old codebases into more modern frameworks in legacy system modernization, traditionally requiring significant human resources and time. "When I was at Shutterstock, we had a huge problem moving from a Perl stack to a modern React stack. The problem was reverse engineering 20 years of code to operate in React...These large language models can figure out how to refactor your code better than a human can." These examples highlight the diverse applicability of generative AI, suggesting that its impact could be far-reaching and transformative across various business operations.

At the heart of Priceline's recent technology initiatives is "Penny" – an AI-powered chatbot designed to enhance the customer experience. This tool, as Marty explains, is not just a generative AI component but is also equipped with reinforced learning and high-performance computing capabilities. Penny represents Priceline's commitment to integrating cutting-edge technology into their services, reflecting its forward-thinking approach to customer interaction and service enhancement. Its design allows it to learn and evolve through real customer interactions, making it increasingly efficient over time. "A lot of the data that feeds Penny and is also feeding other parts of our generative AI capabilities is fundamentally built on getting real-time information from our customers. So that was a very big deal." Marty underscores Penny's iterative development process, where continuous A/B testing and refinements in AI prompts were essential in achieving a tool that positively impacts customer conversion rates. This process, focused on real-time data and customer feedback, ensures that Penny is a technological advancement and a highly effective tool for personalizing customer interactions. Penny represents a significant step forward in how the latest AI tools can revolutionize customer service chatbots.

As CTO, Marty provides invaluable insights and a singular perspective into the strategic implementation of AI in enterprise environments, given his myriad experience at Priceline. While much noise exists about the latest AI capacities, the transformative impact of tools like Penny in enhancing customer experience is a testament to their true potential. As AI continues to evolve, its integration into business strategies will undoubtedly become more nuanced and critical, making discussions like this one essential for understanding the future trajectory of AI in business applications and society.

Listen to Marty's episode here and read the transcript here.