Data & AI on Public Cloud

Problem

Lines of business were adopting cloud data and AI services one design at a time, with inconsistent controls and duplicated effort.

Approach

• Defined a multi-cloud Data & AI strategy with strategic, logical and physical reference architecture, a target state and a phased roadmap.
• Enabled managed AI platforms and the model development lifecycle for predictive and generative models on public cloud.
• Created prescriptive, reusable Cloud Data & AI archetypes so teams could start from an approved pattern.

Governance

• Controlled data movement and enterprise data controls built into every archetype.
• Architecture-led delivery, with every archetype approved once through enterprise architecture review.

Outcome

• Public-cloud onboarding cut from months to weeks using the archetypes and codified patterns.
• One approved path for a growing set of cloud data and AI services across lines of business.
• A legacy data warehouse modernized to a cloud lakehouse at lower infrastructure cost.