Angie Jones on Building Systems, Not Just Code, with Agentic AI

In an insightful presentation from the Agentic AI Foundation, Angie Jones, a prominent figure in software development and AI, laid out a compelling vision for the future of software creation. Her talk, titled "Build Systems, Not Code," argues that the advent of agentic AI is fundamentally reshaping how we approach development, moving the focus from writing explicit lines of code to designing and orchestrating intelligent, autonomous systems. Jones's perspective offers a critical roadmap for developers and organizations navigating the rapidly evolving AI landscape, emphasizing a shift towards higher-level system design rather than granular coding.

Who Is Angie Jones

Angie Jones is a highly respected authority in the software development world, particularly known for her expertise in test automation and her contributions to the developer community. She has held significant roles at major tech companies, including Twilio (NYSE:TWLO) and Apple (NASDAQ:APPLE), where she focused on improving developer experience and software quality. Her work extends beyond traditional coding, often exploring how new technologies can enhance development processes. As a speaker and educator, Jones is adept at breaking down complex technical concepts into understandable insights, making her a valuable voice in discussions about the future of AI and software engineering. Her involvement with the Agentic AI Foundation underscores her commitment to advancing the practical application and understanding of agentic systems.

The Paradigm Shift: From Code to Systems

Jones began her presentation by asserting that the traditional model of software development, heavily reliant on explicit coding, is undergoing a profound transformation. She highlighted that while code has been the bedrock of software for decades, agentic AI introduces a new layer of abstraction and autonomy that demands a different approach.

"We're moving from a world where we write every single instruction to one where we define goals and let intelligent agents figure out the path," Jones explained. This shift implies that developers will spend less time dictating exact steps and more time defining objectives, constraints, and the tools available to an AI agent.

Understanding Agentic AI

The core of Jones's argument rests on the definition and implications of agentic AI. She described agentic systems as those capable of understanding a goal, breaking it down into sub-tasks, selecting appropriate tools or actions, executing those actions, and adapting based on feedback, all without explicit human intervention for each step. This capability moves beyond simple automation; it involves intelligence and decision-making.

Jones emphasized that agentic AI is not just about using large language models (LLMs) but about building entire systems around them. These systems involve orchestrating multiple components: a planner, a memory, tools, and an environment. The planner determines the sequence of actions, memory retains context, tools enable interaction with the outside world, and the environment provides the operational context.

Designing for Autonomy and Goals

A significant portion of Jones's talk focused on the design principles for effective agentic systems. She stressed that the success of such systems hinges on clearly defined goals and well-chosen tools. The goal should be specific enough to guide the agent but broad enough to allow for autonomous problem-solving.

"When you're building an agentic system, you're not coding a solution, you're defining a problem space and providing the means for the agent to navigate it," Jones stated. This means developers must think more like architects and less like coders. They need to anticipate various scenarios, potential tool interactions, and how the agent will recover from failures or unexpected outcomes.

The Role of the Agentic AI Foundation

The Agentic AI Foundation, which Angie Jones represents, plays a crucial role in this evolving landscape. Jones explained that the foundation aims to establish best practices, standards, and educational resources for agentic AI development. The goal is to ensure that as this technology becomes more prevalent, developers have the guidelines and tools necessary to build reliable, safe, and effective agentic systems.

Standardization is key to preventing fragmentation and accelerating adoption. By providing a common framework, the foundation seeks to lower the barrier to entry for developers and organizations looking to integrate agentic capabilities into their products and services.

Implications for Developers and Organizations

For individual developers, Jones's insights suggest a need to evolve skill sets. The focus will shift from mastering specific programming languages or frameworks to understanding system architecture, prompt engineering, and the nuances of AI behavior. Debugging will also change, moving from tracing code execution to understanding why an agent made a particular decision or failed to achieve its goal.

Organizations, on the other hand, must reconsider their development workflows and team structures. Investing in training for system design, AI ethics, and monitoring autonomous systems will become paramount. The promise of agentic AI is increased productivity and the ability to tackle more complex problems, but this requires a strategic shift in how software is conceived and delivered.

Jones concluded by reiterating that the future of software development is not about eliminating code entirely but about elevating the developer's role to a higher level of abstraction. By building systems that can intelligently interpret goals and execute tasks, we unlock a new era of software that is more powerful, adaptable, and efficient.