AI Enterprise - Databricks & Glean | BG2 Guest Interview - BG2Pod with Brad Gerstner and Bill Gurley Recap

Podcast: BG2Pod with Brad Gerstner and Bill Gurley

Published: 2025-12-23

Duration: 45 min

Summary

The episode delves into the current state of AI, discussing its commoditization, the ongoing bubble in the industry, and successful use cases across various sectors like finance, healthcare, and retail.

What Happened

In this episode of BG2Pod, Brad Gerstner and Bill Gurley explore the complex landscape of artificial intelligence, particularly focusing on the divide between consumer and enterprise applications. They highlight the stark reality where, despite reports indicating that 95% of AI projects fail, this isn't necessarily a bad sign. It suggests that organizations are experimenting with new technologies, which is crucial for progress. The discussion emphasizes that successful AI implementations require more than just the technology; they need a solid understanding of unique company data and a skilled team to bring these projects to fruition.

The conversation shifts to specific success stories in various industries, illustrating how companies like the Royal Bank of Canada and Merck are leveraging AI to transform their operations. The RBC has developed agents that can analyze earnings reports and generate comprehensive equity reports significantly faster than traditional methods. In healthcare, Merck’s TEDDI model is aiding drug discovery by predicting gene expressions, showcasing AI's potential to revolutionize fields that rely heavily on data. The episode underscores that while many AI projects are unsuccessful, the successful ones are paving the way for tangible benefits in business.

Gurley and Gerstner further discuss the commoditization of large language models (LLMs), likening them to interchangeable products where price becomes the primary differentiator. They argue that to thrive, companies must focus on leveraging their unique data assets and building AI solutions that truly understand their specific business contexts. This realization is essential as organizations navigate the current AI bubble, distinguishing between mere hype and real value generation. Overall, the episode provides an insightful overview of the state of AI, urging listeners to embrace experimentation while being mindful of the challenges ahead.

Key Insights

• 95% of AI project failures indicate a healthy experimentation culture rather than a lack of potential.
• Successful AI implementations require unique company data and skilled teams to leverage that data effectively.
• The commoditization of LLMs means businesses need to focus on their unique data to gain a competitive edge.
• Real-world use cases in finance, healthcare, and retail demonstrate the transformative potential of AI when applied correctly.

Key Questions Answered

What does the 95% failure rate of AI projects mean?

Irving challenges the common perception surrounding the 95% failure rate of AI projects, framing it as a necessary aspect of experimentation. He argues that if all projects were successful, it would indicate a lack of innovation and willingness to explore new technologies. By failing, companies are actively trying to understand and integrate AI into their operations, which is essential for progress and eventual success.

How are companies successfully implementing AI in finance?

One notable example shared by Irvind is the Royal Bank of Canada's development of agents that automate the equity research process. These agents quickly analyze earnings reports, pulling data from various sources to generate comprehensive equity reports in just 15 minutes, a significant reduction from the two-hour industry standard. This demonstrates a successful application of AI that not only saves time but also transforms the workflow in finance.

What role does unique company data play in AI success?

Irving emphasizes that while LLMs are now commodities, the real differentiator for companies lies in their unique data. He stresses the importance of leveraging proprietary data to build AI systems that understand specific business processes. This tailored approach is what will allow companies to create competitive advantages, as generic AI solutions cannot replicate the insights derived from unique datasets.

What examples of AI success are there in healthcare?

In the healthcare sector, Merck's TEDDI model stands out. This AI model is designed for drug discovery, capable of predicting gene interactions and understanding gene regulatory networks. Such advancements highlight the potential of AI to significantly impact fields like pharmaceuticals, enabling breakthroughs that were previously unattainable.

How is AI transforming marketing in retail?

Irving points to 7-Eleven's use of AI agents to automate their marketing stack as a prime example of AI's transformative potential in retail. These agents can segment audiences, prepare targeted marketing materials, and create customized campaigns, which streamlines marketing efforts and enhances customer engagement. This shift shows how AI can reduce the manual labor typically associated with content creation in marketing.