Jon Farr

Semi-structured data remains a hidden limiter of enterprise AI. Decades of unavoidable architectural tradeoffs refining JSON and XML-based standards like FHIR and ACORD introduced blind spots and technical debt that models can’t reason over. This executive roundtable examines why that constraint endured, what has changed by design, and includes a customer perspective from a company that overcame it, resulting in a seven-figure cost reduction.

Learn how C.H. Robinson used DataPancake (a Snowflake Native App) to migrate their Databricks solution to Snowflake. In record time, C.H. Robinson created accurate and secure data pipelines using Dynamic Tables and policy-infused views to flatten complex JSON from Kafka streams. DataPancake also enabled C.H. Robinson to create a Cortex AI-enhanced data glossary and Cortex Analyst semantic models, resulting in cost savings, reduced engineering time, and increased data availability and quality.