At Workiva Amplify, theCUBE Panel Says AI Governance Must Set Limits on Agent Autonomy

Krista Case, an analyst at theCUBE Research, said on an exclusive broadcast that organizations deploying agents in finance, audit, compliance and sustainability need results grounded in business context, traceable to their sources and covered by clear accountability. Case and co-host Alison Kosik spoke with executives and practitioners at the event for theCUBE, SiliconANGLE Media's livestreaming studio. "We need to understand things like: Where did this information and … these insights come from? Who approved a particular action? Can you maybe trace what happened if an AI agent is taking an action on your behalf?" Case said. "When you think about finance and these other regulated processes, it's not really good enough for the action or the response to just look plausible. We have to really make sure that it can be substantiated."

The harder design question, she said, is how much independence to grant an agent without giving up meaningful oversight. If a person must review and approve every action, much of the value of automation disappears, Case said, so companies have to decide when AI assists a human, when it can execute on its own and when it requires approval. "I do think that some of these boundaries are still being defined, and I think they're going to evolve and change over time, especially as the business use cases evolve," she said.

Workiva's 2026 Midyear Executive Benchmark Survey, cited in the coverage, points to a gap between confidence and readiness. Eighty-four percent of surveyed executives were at least somewhat confident in AI outputs appearing in annual reports without human review, while only 11 percent said their data quality was sufficient for AI use.

Steve Soter, vice president and industry principal at Workiva, said established financial controls become more important as AI accelerates reporting. Those controls document four elements: where data originated, who touched it, what happened to it, and how it was reviewed and approved. "It actually makes it even more important because accelerating a process, if you don't have it grounded by … those four things, then speed doesn't become an asset," he said. "It really becomes a liability. It becomes a risk."

Case described Workiva's larger opportunity as creating the context that connects disparate platforms and brings in institutional or "tribal" knowledge, so that agents asked to execute functions can be trusted because they have that context behind them. "They are historically very strong in things like compliance and reporting and very regulated and audited workflows," she said. "They are beginning to make some steps into areas like sustainability."

Workiva Chief Information Officer Kim Huffman said Agent Studio gives business users a no-code way to build and customize agents for their own processes, and that early teams produced useful agents within days. The company applies an outcome-first approach internally: employees across functions define a shared result, then determine where agents can perform parts of the work. Huffman described AI as a fundamental technology that will reshape the landscape.