Agency-based artificial intelligence (AI) is transforming how businesses develop and use software. Traditionally, business experts—like supply chain managers or financial analysts—have identified inefficiencies and needed tools, but they often rely on the IT department to build these solutions. This creates a disconnect: the people who understand the work best rarely have the tools to create the systems that could improve it, while those who can build the systems are not deeply involved in daily operations. As a result, ideas can get stuck in long queues, especially if they seem small or low priority. Agency-based applications aim to bridge this gap by allowing those who know the work best to design solutions while keeping them secure and aligned with the company's existing systems and policies. The shift is already happening. According to Gartner, 80% of low-code platform users are now from departments outside of IT, up from 60% in 2021. This trend highlights a growing movement where the people who use enterprise software are also the ones developing it. However, this transition comes with challenges. AI tools are often built in isolation, separate from the company’s central systems. This leads to an additional, often overlooked, project: integrating essential controls like security, permissions, and traceability after the fact. Many pilot projects fail at this stage because these critical elements are not considered early on. To address this, a new generation of enterprise technologies is emerging, supporting what some call a "builder economy." This concept involves expanding the number of employees who can create applications, provided they do so securely and within the company's governance framework. Some platforms now allow business users, developers, and integrators to build AI-based applications directly, with security and governance built in from the start. This approach ensures that everyone can use the tools they are familiar with, whether through natural language or traditional coding, without compromising the company's safety or control systems. Agency-based applications represent a new kind of enterprise software, designed to focus on achieving results rather than just recording data. These applications use teams of specialized agents that can understand the company's current state, reason through problems, coordinate actions, and make decisions. Unlike isolated automation tools, they work directly within the systems where work is already happening, ensuring that decisions are made with the right context and safeguards. As a result, long-standing recording systems can evolve into true results systems, improving financial closure times, reducing service escalations, and optimizing staff planning. Organizations that have successfully moved AI from experimental stages to widespread use are seeing significant returns, with cost reductions in supply chain, finance, and operations. The future of enterprise software will likely depend on how quickly organizations can develop their own secure solutions, rather than relying solely on external publishers.