Why data teams trail developers in agentic AI gains
The article argues that data workers have lagged behind software developers in gaining productivity from agentic AI because their results are harder to inspect and validate. It outlines three solvable causes, starting with business users' difficulty accessing governed data and the resulting reliance on stale or copied datasets. This weakens data lineage and makes it hard to trust AI-generated analyses, even when the tools produce answers quickly.
Software engineers adopted agentic AI earlier partly because their work can be inspected: they review changes and run tests before accepting results. Business and financial operations employees, who outnumber U.S. software developers roughly six to one, often receive only figures, charts, or code they cannot readily verify.
Three fixable issues hold them back. Governed data is hard for business users to reach, so they rely on outdated copies and lose lineage. There is no practical way to fact-check AI outputs. Many tools also presume clean, table-ready data and skip preparation, limiting real-world use.
The gap may shape how quickly non-engineering teams adopt agentic AI. Business and financial operations staff could gain speed but may act on analyses they cannot fully audit, affecting decisions, reporting, and accountability. Auditors and managers may face harder questions about provenance. If vendors address data access, verification, and preparation, trust could improve; if not, productivity gains may remain concentrated among developers and data platform teams.