Why agents are the new kingmakers

In 2013, Redmonk analyst Steven O’Grady positioned application developers as the new kingmakers. It was a role that enterprise IT had served since the rise of business-driven computing. First, systems administrators held the keys to the kingdom by having the (then) esoteric knowledge of the operating system - but as Linux took hold in the late 90s/early 00s, the applications, not the OS, took center stage. 

This made developers the unlikely “voice behind the throne” in a CxO monarchy. But we’re looking at another shift in royalty fabrication with the continued velocity of generative AI (gen AI) and AI-driven automation.

Specifically, agents, essentially autonomous AI-derived avatars trained by developers, are quietly becoming the *new* new kingmakers in the world of business IT. According to Boston Consulting Group, AI agents are growing across tech applications, seeing a compound annual growth rate of nearly 46% over the next five years (from $5.7 billion in 2024 to $52.1 billion in 2030).

But if humans aren’t really needed in determining the new successful business (monarchy)...what happens next?

The rise of agents

Traditionally, developers have been the architects of software, meticulously crafting each line of code, hand selecting open source frameworks and libraries, and then putting it all together on their favorite platform (Linux, Kubernetes, serverless, etc). However, with the advent of agents, intelligent systems that can autonomously handle routine tasks will do more and more of this work.  In many cases these agents will select the libraries, frameworks, and platforms in addition to writing custom code . 

Many AI coding assistants have context files, a concept similar to a bashrc file, where developers can specify prepended, standardized prompts for their projects. They can add things like code style guidance, what tech stack to use, what directories to use, how to test code and more.  Then, when the engineer types a new prompt, the model will use standardized context for everything it generates. From that step forward, the model will do the actual work without direct human intervention. If the engineer doesn’t specify a platform or tech stack, the agent is likely to specify one for them!

These agents are not just tools; they are collaborators. They can analyze vast amounts of data, make informed decisions, and execute tasks with precision. In essence, they are becoming the unseen hands that are poised to drive technology decisions forward.

An enterprise developer in King AI’s court

Given the raft of decisions that AI agents will make in the foundational building of an application, it’s easy to have a minor panic attack about humanity ceding this role to a machine. The decisions made by agents shape development lifecycles, influence code quality and deployment and even drive platform selection. 

But this power shift still has a human being at its core.I

We, as humans, determine what data the agents are trained on, how tasks are prioritized, which processes are automated, and how resources are allocated. It’s also up to us, the humans, to influence the data and make our open source technology stacks visible and usable to these agents.

The impact on developers

For developers, this evolution presents both opportunities and challenges. On one hand, agents can alleviate the burden of repetitive tasks, allowing developers to focus on more creative and impactful work. On the other hand, the reliance on these systems means that developers must understand how to design, implement, and maintain them effectively.

As agents become more integrated into the development process, they may begin to influence decisions traditionally made by developers, or completely make the decisions for them. For instance, an agent might determine the optimal time to deploy code based on historical data, or what tech stack to use, potentially surprising developers. This surely raises questions about autonomy and control in the development process.

Preparing for the Future

To thrive in this new paradigm, developers must adapt. Understanding the principles of agent design and operation will become as fundamental as mastering programming languages. Collaborating with agents, rather than viewing them as mere tools, will be key to unlocking their full potential.

At Red Hat, we are committed to empowering developers with the tools and knowledge they need to navigate this evolving landscape. We’re also committed to helping open source communities understand how this new paradigm will affect their projects, especially the consumption of their projects by these agents. By embracing, rather than resisting, the rise of agents, developers can position themselves at the forefront of the next wave in software development and ultimately leave their mark on the automation driving application innovation forward.

Agents, as the technology stands today, are inevitable in how they will continue to evolve and integrate into development workflows. These are not just supporting actors—they are becoming the new kingmakers. Those who understand and harness their power will lead the charge in shaping the future of software development - so maybe we should start looking at developers as the crafters of kingmakers?