Agentic AI in the Enterprise Part 2: Guidance by Persona
AWS details how business owners, CTOs, and security leaders must collaborate to scale AI agents beyond pilot projects.
The AWS Generative AI Innovation Center has released the second part of its practical guide for enterprises looking to deploy agentic AI—autonomous AI systems that can take actions across tools and systems. Titled 'Agentic AI in the Enterprise Part 2: Guidance by Persona,' the report argues that the primary barrier to success is not the technology itself, but the operating model. It provides targeted advice for different leadership roles, starting with the line-of-business owner who must treat an AI agent like a new hire, writing a precise job description tied directly to operational KPIs like ticket resolution time or cash conversion cycles.
For the CTO or chief architect, the guide warns of the risk of uncontrolled success, where early wins lead to a proliferation of incompatible, one-off agents. It advocates for building a standardized system from the outset that can safely support hundreds of agents, not just ten. This requires separating an agent's 'thinking' from its 'doing,' standardizing how tools and data are exposed, and designing agents as long-lived services with proper identities, permissions, and lifecycle management. The underlying message is that scalable, valuable agentic AI requires deliberate architectural and governance decisions made early, long before the first agent is deployed.
- Line-of-business owners must write a formal 'job description' for an AI agent, anchoring its business case in existing operational metrics like cost per unit or queue wait times.
- CTOs must choose between building impressive one-off agents or a standardized system capable of supporting hundreds; the latter requires early investment in unified tool integrations and agent observability.
- The biggest barrier to agentic AI is the operating model, not the technology; success depends on defining work precisely, bounding autonomy, and treating improvement as a continuous habit.
Why It Matters
Provides a concrete playbook for executives to move AI agents from experimental pilots to scalable, governed systems that directly impact business metrics.