Microsoft is developing an agentic software stack through its CoreAI division to fundamentally change how enterprise AI applications are run and secured [1].
This shift represents a departure from traditional software engineering. As AI capabilities evolve rapidly, the company is treating CoreAI as an internal startup to maintain the agility needed to build a scalable infrastructure for AI agents.
Jay Parikh, executive vice president of CoreAI at Microsoft, said that the pace of change has rendered traditional long-term planning obsolete. He said that roadmaps designed for six months or a year can become outdated within a few weeks [1].
Parikh said the current environment is a complete reset in software construction. The new agentic stack is designed to manage the complexities of AI at an enterprise scale, focusing on the orchestration, management, and security of these autonomous systems [1].
While Microsoft focuses on infrastructure, other industry players are restructuring to adapt to the same trend. GitLab is reorganizing its research and development into 60 autonomous teams [2]. The company is also reducing its country footprint by 30 percent [2].
Bill Staples said these types of industry shifts are an investment in the agentic era rather than a simple cost-cut [2]. This reflects a broader movement across the tech sector to move away from static software toward dynamic, agent-driven ecosystems.
“You can make a plan for a roadmap for six months or a year and within a few weeks it's going to be out of date.”
The transition toward an 'agentic' software stack signals a move from software that merely follows instructions to software that can independently execute complex workflows. For enterprises, this means the traditional software development lifecycle is being replaced by a more fluid, iterative process where the infrastructure must be capable of updating in real-time to keep pace with AI model advancements.
