Experts Urge Companies to Start Small With Generative AI
Tech industry leaders advise businesses to avoid data overload by focusing on specific, small-scale goals when adopting generative AI. Starting with internal applications helps prevent expensive and inaccurate outcomes.
Industry leaders at TechCrunch Disrupt 2024 warn that companies often become overwhelmed by data when adopting generative AI. DataStax CEO Chet Kapoor emphasizes that while unstructured data at scale is essential for AI, the sheer volume of information across various locations creates a significant hurdle for businesses. Instead of attempting to tackle everything at once, experts advise organizations to rely on small, specialized teams to navigate these early days of AI development.
Vanessa Larco, a partner at NEA, recommends a pragmatic approach where companies work backwards from their specific business goals. Rather than throwing all corporate data at a large language model and hoping for the best, businesses identify exactly what problem they need to solve and locate only the specific data required to address it. Attempting to deploy generative AI across an entire company from the outset usually results in an expensive and inaccurate mess.
The consensus among the panelists is to start small with internal applications that have highly targeted objectives. By focusing on practical, incremental progress rather than massive scale, companies effectively write the manual for future AI implementations. This careful, deliberate strategy allows organizations to unlock true value from generative AI without falling victim to data overload.