Databricks scientist: Companies finding AI value through experimentation, not buzzwords

Databricks‘s chief AI scientist reveals that companies are still learning through experimentation where artificial intelligence can most effectively solve business problems, beyond the hype of buzzwords like “agentic AI.” This trial-and-error approach highlights a practical shift in enterprise AI adoption, as organizations seek to identify the specific applications where AI delivers meaningful value rather than pursuing trendy concepts.

The big picture: Generative AI is enabling a new form of enterprise analytics by making previously unusable unstructured data valuable for business insights.

• Data that was once considered “useless” in traditional analytics frameworks—like Word documents, images, and videos—has become “incredibly valuable” with the emergence of large language models.
• This transformation represents a fundamental shift in how companies can extract meaningful features and insights from their vast stores of unstructured information.

Why this matters: Companies are moving beyond AI buzzwords to discover practical applications through hands-on experimentation rather than theoretical implementations.

• The focus has shifted from chasing trending concepts to identifying specific use cases where AI technology can demonstrably solve real business problems.
• This pragmatic approach reflects the maturing enterprise AI market, where value is increasingly measured by practical outcomes rather than technological novelty.

Key details: Databricks is witnessing firsthand how businesses are systematically testing AI applications to find the optimal use cases.

• Jonathan Frankle, Databricks’s chief AI scientist, emphasized that companies are actively searching for the “sweet spot” where AI provides genuine problem-solving capabilities.
• The company’s perspective comes from its position as a major data tools provider with visibility across numerous enterprise AI implementations.

Reading between the lines: The industry appears to be entering a more practical phase of AI adoption, prioritizing tangible results over speculative implementations.

• This signals a potential shift away from the investment-driving hype cycle toward more sustainable, results-oriented AI deployment strategies.
• Companies may increasingly demand demonstrable ROI before committing to large-scale AI projects, regardless of how trendy the underlying technology might be.