COOs must rework end-to-end workflows and redefine human‑agent responsibilities to capture the value of agentic AI, rather than merely automating existing tasks. Deploying AI agents into current roles and processes without redesigning the work itself is unlikely to deliver the expected benefits; success depends on shifting from fixed task completion to fluid human‑agent collaboration and rebuilding operating models with clear accountability at every step.
Agentic AI and the need to redesign workflows
Approaching process change as a redesign challenge—not an optimization exercise—allows humans and agents to play to their strengths while embedding accountability measures into each stage. Alessio Marras, cofounder and head of organization for AideXa, an Italian digital bank serving small- and medium-sized businesses, says: “We’re developing a broader strategy. It’s based on a clear vision for where and how we want to use AI agents, then implementing them progressively. The challenge is coordination–moving from one use case to multiple orchestrated agents requires an end-to-end view. Agents that are very efficient in single use cases can become much more difficult to coordinate and govern when they operate as part of an orchestrated system. That’s why we believe in continuously monitoring outcomes and being ready to adjust the balance between automation and human oversight as processes evolve.”
AI can reduce the time employees spend on execution, analysis and routine tasks, freeing them to focus on strategic, value‑creating priorities. In the near term however, organizations may struggle to build the internal agentic AI expertise needed to scale. Employees who have collaborated with external teams may need reskilling to operate effectively in an agent‑based environment.
Managing accountability, risk and scaling agent networks
COOs and senior leaders must manage accountability, risk and performance carefully, establishing guardrails that account for multi‑agent systems and the dynamic relationships between humans and machines. As agent networks expand, coordination complexity can grow rapidly across business units, systems and decision points. Leaders should define how work flows between humans and agents, locate escalation points, monitor performance across the full system rather than within isolated use cases, and preserve visibility and control as the network scales. Getting these orchestration layers right will likely distinguish organizations that build coherent, enterprise‑wide operational capabilities from those that implement only a collection of AI experiments.
