What is most critical for scaling AI enterprise-wide?

Anahita Tafvizi

Chief Data & AI Officer · Snowflake

Unified data governance

Anahita Tafvizi says the key requirement is unified data governance and unified metric definitions, arguing that without that foundation AI just scales a company's chaos rather than giving trusted answers.

BH

Beth Hiatt

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

AI data infrastructure blueprint

Beth Hiatt says companies should have an AI data infrastructure and control plane blueprint in place within the first 90 days, because she thinks most companies underestimate the work that stack requires and having it in place is what makes AI initiatives successful.

The answer starts at 5:20.

CP

Conny Ploth

AI Transformation Vice President, Global Data & AI · Santander

Getting data in order

Conny Ploth says getting data in order is critical, since AI is only as good as the data underneath, and without well-structured data and identified sources you won't get the right results.

The answer starts at 5:36.

Sam Pierson

Chief Technology Officer · Qlik

Reliable high‑quality data

Sam Pierson says a reliable, high-quality data foundation is number one, along with context on top, because it raises the quality and accuracy of LLM answers.

Kevin Lee

Chief Technology Officer & Key Pursuits Leader · Nice

Compliance rigor

Kevin Lee says compliance rigor is critical for scaling AI past pilots, because organizations in regulated industries that lean into requirements like HIPAA and PII, rather than shy away, are yielding the best results.

Muralidhar Vemulapalli

Chief Enterprise Architect/ Acting CTO · Arkansas Blue Cross And Blue Shield

Governance

Muralidhar Vemulapalli says governance is key, arguing that governance is what makes innovation scalable and trusted and that ungoverned AI is not deployable.

Shahir Daya

Chief Product & Technology Officer · Zafin

Workflow-level controls

Shahir Daya says defining controls at the workflow level, not the agent level, is the first move for enterprise-grade agents, since one agent can work across many workflows and one workflow can be run by many agents.