AI governance moves closer to the workflow: theCUBE Insights at Amplify

AI governance is evolving to encompass systems that perform work, not just assist employees. This shift is particularly relevant for companies like Workiva Inc., which deal with critical reporting, audit, and compliance workflows where errors could have severe consequences. When introducing AI agents into finance and other regulated areas, businesses must not only ensure plausible output but also be able to trace information, identify approvals, and reconstruct agent actions, according to Krista Case, principal analyst at theCUBE Research.

In an exclusive broadcast at Workiva’s Amplify event, Case discussed with host Alison Kosik how companies can harness AI for consequential work while maintaining trust, accountability, and control. While AI can process information rapidly across multiple workflows, the quality of the underlying data remains crucial. Fragmented data stores, inconsistent definitions, and unclear data ownership were present before generative AI, but these issues become more urgent when AI agents can transform flawed information into decisions at a faster pace and larger scale.

Case emphasized that AI's effectiveness hinges on the data it is built on. Nonetheless, determining when an agent should act independently versus requiring human approval is a significant challenge. While stringent human oversight might negate the efficiency gains from automation, there is a need for risk-based controls, enhanced monitoring, and exception handling as AI use cases expand.

As business needs evolve, the boundaries around AI involvement are continually being defined and refined.