In an agent-first enterprise, AI systems drive processes while humans set goals, define policy constraints, and handle exceptions.
“You need to shift the operating model to humans as governors and agents as operators,” says Scott Rodgers, global chief architect and US CTO of Deloitte Microsoft Technology Practice.
agent-first imperative
Technology budgets for AI are expected to grow by more than 70% over the next two years, with AI agents, powered by generic AI, poised to fundamentally transform organizations and achieve results beyond traditional automation. These initiatives have the potential to generate significant performance gains while shifting humans toward higher value work.
AI is advancing so rapidly that a static approach to task automation will likely yield only incremental benefits. Since legacy processes are not built for autonomous systems, according to Rodgers, AI agents require machine-readable process definitions, clear policy constraints, and structured data flows.
Making matters more complicated, many organizations do not understand the full economic drivers of their business, such as cost of service and cost per transaction. As a result, they have trouble prioritizing the agents who can generate the most value and instead focus on attractive pilots. To achieve structural change, officials must think differently.
Instead companies must organize results faster than competitors. “The real risk is not that AI won’t work — it’s that competitors will redesign their operating models while you’re still piloting agents and co-pilots,” says Rodgers. “Nonlinear benefits accrue when companies create agent-centric workflows with human administration and adaptive orchestration.”
Routine and repetitive tasks are increasingly being handled automatically, freeing up employees to focus on higher value, creative and strategic tasks. This transformation improves operational efficiency, fosters stronger collaboration, and enables faster decision making—helping organizations modernize the workplace without sacrificing enterprise security.
This content was produced by Insights, the custom content arm of MIT Technology Review. It was not written by the editorial staff of MIT Technology Review. It was researched, designed, and written by human writers, editors, analysts, and illustrators. This includes writing surveys and collecting data for the surveys. The AI ​​tools that may have been used were limited to secondary production processes that underwent thorough human review.
