What practice best ensures governance for AI platforms?

Kumar Veetrag

Chief Technology Officer · Enparadigm

Establish guardrails

Kumar Veetrag says that dealing with AI requires providing enough guardrails and configuring it rather than just judging results, because AI shouldn't replace human judgment but improve its quality.

BH

Beth Hiatt

SVP|Product Management|Data Governance and Responsible AI @LPL Financial · Lpl Financial

Blueprint

Beth Hiatt says companies should have a blueprint for AI data infrastructure and a control plane in place within the first 90 days before branching into AI use cases, because they underestimate that entire stack and the controls needed.

The answer starts at 5:20.

John F Donovan

Principal Research Director · Info-Tech Research Group

Inspect and refine governance settings

John F Donovan says inspecting and refining governance settings before installing Copilot makes risks like PII exposure, IP leakage and morale problems largely avoidable, because AI amplifies existing governance weaknesses.

Mandy Andress

CISO · Elastic

Inventory AI usage

Mandy Andress says the first priority is to get an inventory of AI usage across the organization, because without knowing where it is used and for what objectives, it is harder to put controls and structure around it.

The answer starts at 15:01.

Robert Dickens

Forward Deployed AI Engineer · Cloudera

Explainability

Robert Dickens says explainability of AI matters from the outset: every source must be cited so analysts can trace summarized answers back to original material, which builds trust and helps reduce hallucinations.

Shahir Daya

Chief Product & Technology Officer · Zafin

Define controls at the workflow level

Shahir Daya says the first move is to define controls at the workflow level rather than the agent level, because regulators focus on workflows and agent-level controls won't scale as agents and workflows overlap many-to-many.