What practice helps AI pilots succeed in production?

CC

Chris Carmody

Chief Information Officer · Upmc

Governance

Chris Carmody says it starts with governance and the processes that support it, so an organization can identify the few problems it wants to solve, focus its efforts there and build off those successes.

The answer starts at 7:04.

Praveen Jayakumar

Head of Technology Generative and Agentic AI, Asia-Pacific & Japan · Amazon Web Services (Aws)

Observability

Praveen Jayakumar says one thing he tells customers is to get observability right in the proof of concept itself, not as an afterthought, so they can go into one place to see what is failing.

The answer starts at 7:11.

Scot Burdette

Global Division CIO, Measurement and Analytics · Abb

Design, Develop and Implement

Scot Burdette says the real work after a pilot is to design, develop and implement it, because that's where transformation happens.

Julian Tang

COO for Innovation Office · BlackRock

Proactive Governance

Julian Tang says proactive governance — clear playbooks, transparent communication and early involvement from risk, legal and infosec — is what eliminates shadow AI and creates the conditions for intentional, repeatable AI deployment.

Pete Johnson

Field CTO, Artificial Intelligence · MongoDB

Pick the Right Problem

Pete Johnson says the key is picking the right problem, because you need good data for it and metrics to measure success.

Kevin Lee

Chief Technology Officer & Key Pursuits Leader · Nice

Choosing the Right Partner

Kevin Lee says organizations should be more discerning in choosing the right partner, one of two things to focus on at the outset, because a partner lacking compliance rigor and enterprise scale will cause the pilot to stall.