AI Budget Planning Starts With Infrastructure
AI infrastructure readiness is becoming a core budget issue for CIOs as they prepare for 2027. Jack Lodge, Executive Vice President of Customer Success at NWN, explains why many organizations cannot simply add agentic AI on top of existing environments. The models and agents may be powerful, but many desktops, networks and data center systems still carry technical debt.
That gap is shaping refresh plans across the enterprise. AI infrastructure readiness requires better endpoint performance, stronger network design and more thoughtful cloud and data center strategies. Lodge notes that AI PCs with NPUs can reduce network strain by processing more AI workloads locally. At the same time, enterprise networks need to support more east-west and agent-to-agent traffic.

Agentic AI Raises New Security Questions
The discussion also turns to governance and identity. AI agents may inherit user permissions, interact with other agents and operate semi-autonomously. That makes them different from both human users and traditional non-human identities. Security teams need to know which agents exist, what they can access and whether their actions are sanctioned.
Lodge argues that agent identity management will become a top 2027 investment priority. Enterprises want the productivity gains promised by agentic workflows. They also need guardrails, logging and auditability so agents do not create more risk than value.

Refresh Cycles Can Help Fund AI
Most organizations will not receive unlimited AI budgets. Lodge says IT leaders need to drive efficiency in legacy environments and use those savings to fund new investments. That can include retiring run-rate spending, modernizing endpoints and rethinking GPU access as a sourcing strategy rather than only a capital expense.
The refresh cycle gives CIOs a practical path forward. Instead of upgrading PCs and networks for their own sake, teams can align those investments with AI transformation. That makes infrastructure modernization part of the business case for agentic AI.

Managed Services Support a More Horizontal IT Model
The conversation also explores whether more infrastructure operations will shift to managed services. Lodge says many customers want providers such as NWN to operate foundational environments. That lets internal teams focus more of their time on agentic workflows, business outcomes and transformation work.
AI also challenges the traditional IT silo model. Endpoint, network, data center, cloud, communications and security teams can no longer operate as separate islands. AI infrastructure readiness depends on visibility across the entire user experience. It also requires insight into how each layer affects business outcomes. For Lodge, the opportunity is to move from legacy thinking to a more horizontal model that supports AI-enabled operations.