Synopsis: A successful AI pilot proves less than many organizations expect. Scaling it means preparing people, changing processes and deciding who owns the results. Adding a model to an existing workflow may improve productivity, but larger gains can require redesigning the work itself. The gap between buying technology and building the capability to use it helps explain why investment can arrive well before measurable business value.
EY Distinguished Technologist Rick Ross examines that gap with Mike Vizard, arguing that organizational readiness deserves as much attention as the technology. He recommends starting with a specific process or capability and identifying the value a change is meant to create. That gives teams a narrower, more useful question than whether AI is delivering a return across the entire enterprise. Investment strategies will differ with resources and risk appetite, but training people and preparing operating models cannot be left until after deployment.
Software agents make those operating-model changes concrete. Ross describes agents requesting services, interacting with other systems and moving through their own onboarding, role changes and retirement. Those activities need lifecycle management and oversight, just as employee access and responsibilities do. He also raises the need for independent validation as organizations accelerate development or build capabilities themselves. Enterprise and solution architects can help challenge assumptions and provide quality controls rather than leaving the builders to judge their own work.
Protecting intellectual property adds an architectural decision to the adoption plan. Ross outlines a mixed approach in which externally hosted models handle suitable tasks while selected internal workloads and sensitive information remain in controlled environments with local processing. The right boundary depends on what the organization needs to protect and where it expects to gain an advantage. His advice combines action with preparation: safeguard important assets, invest where there is a credible opportunity and develop the skills needed to support that investment. Moving quickly is not a substitute for knowing who will operate, govern and maintain what gets built.
