GenAI suffers from data overload, so companies should focus on smaller, specific goals

At TechCrunch Disrupt 2024, Chet Kapoor, the chairman and CEO of DataStax, highlighted the foundational role of data in artificial intelligence (AI), asserting that AI’s success hinges on structured and unstructured data, especially unstructured data at scale. This assertion was part of a larger discussion involving Vanessa Larco, a partner at NEA, and George Fraser, CEO of Fivetran, centered around the significance of “new data pipelines” in contemporary AI applications.

A key theme of their conversation revolved around the importance of balancing ambition with practicality, especially within the early stages of AI development. Kapoor emphasized that generative AI’s true success relies not just on technical resources but on the people behind the technology, suggesting that the pioneering teams are essentially creating the guidelines for future projects in AI.

While data plays a crucial role, companies often face the challenge of managing and extracting value from vast amounts of distributed and potentially sensitive data. Larco offered a pragmatic approach to navigating this complex landscape—advocating for a reverse-engineering strategy that begins with the specific problems businesses aim to solve. Instead of adopting a broad, unfocused application of generative AI across entire organizations, Larco encouraged companies to start with targeted applications that accurately align with their objectives. By identifying relevant data and utilizing it effectively for specific purposes, firms can avoid the pitfalls of misapplication and inefficiency that tend to occur when attempting to forcefit generative AI technologies generically across their operations.

Fraser reinforced the notion of adopting a narrow focus by advising firms to address immediate problems without aiming for grand-scale implementations prematurely. He articulated a mantra of “only solve the problems you have today,” cautioning against the often-overlooked costs associated with failed innovations, which can dominate the expenditure landscape more than successful projects.

Drawing parallels with the early expansion of the internet and the smartphone revolution, Kapoor categorized the current phase of generative AI development as the “Angry Birds era.” While various applications have emerged, including internal tools and smaller-scale productions, none have fundamentally transformed everyday life yet. However, Kapoor noted that enterprises are beginning to implement small-scale applications and refine their approaches as they experiment with generative AI solutions. He optimistically predicted that the next year would mark a pivotal turning point for companies, ushering in genuine transformation driven by more impactful AI applications.

Overall, the discussion underscored an emerging consensus that organizations venturing into the realm of generative AI should prioritize understanding their specific needs and leverage data strategically, while being mindful of the ongoing development and evolutionary nature of AI technology. Companies are urged to adopt a gradual approach, focusing on iterative learning and practical implementations rather than expansive and untested ambitions. This mindset not only fosters innovation but also positions businesses better for long-term, meaningful advancements in AI functionalities, ultimately shaping a more efficient and data-driven future.