Matilda Bailey has spent her career dissecting the intricate layers of next-generation networking, but at this year’s Dell Technologies World conference, the conversation shifted from the pipes to the intelligence flowing through them. As a specialist in cellular and wireless solutions, she understands that the true value of any network lies in how effectively it supports the work of the enterprise. The emergence of agentic AI represents a fundamental leap in this evolution, moving us away from the era of simple queries and toward a future of autonomous reasoning and digital execution. In this discussion, we explore the nuances of this technology, the necessity of a protocol-based approach to ethics, and why the speed of an organization’s execution has become the most critical differentiator in a world where software code is increasingly commoditized. We delve into how businesses can balance the urgent need for innovation with the stringent requirements of security, ensuring that autonomous agents become a source of strength rather than a liability.
Agentic AI is often mischaracterized as just another layer of chatbot technology, yet the underlying architecture suggests something far more profound. How should we redefine our understanding of these autonomous entities to better reflect their role in the modern workforce?
To truly grasp the power of agentic AI, we have to look past the chat window and see it as the digitization of work itself rather than just a sophisticated productivity tool. We are talking about autonomous entities capable of reasoning, using tools, maintaining memory, and possessing specialized knowledge to solve complex problems. In the conference session titled “Agentic AI: Unlocking the Power of Purposeful Intelligence,” it was made clear that these agents are not just answering questions; they are interacting with other agents to optimize workflows and fulfill specific objectives. This isn’t about a human typing a prompt and getting a summary; it’s about an agent seeing a task, reasoning through the steps required to complete it, and then executing those actions independently. When you witness an agent autonomously navigating through multiple enterprise tools to resolve a supply chain bottleneck, you feel the sensory shift from a “helper” to a “doer,” which fundamentally changes the human and organizational dynamics of the company.
The categorization of agents—ranging from low-autonomy tools to sophisticated coordination agents—is a specific framework used to guide deployment. How does this taxonomy provide the necessary guardrails for an organization looking to achieve a measurable return on investment?
Establishing clear guardrails allows a business to correlate the technology directly to the specific thing it is doing, which is the only way to see results faster and avoid the “random adventure” approach. The framework identifies four operational functions: simple productivity tools with low autonomy, entities that handle simple work with high autonomy, hygiene agents that fulfill broad objectives with many tools, and coordination agents that monitor complex workflows. By being deliberate about the scope and intentions of these agents, a company can save significant time and money while producing results with remarkable efficacy. For example, a hygiene agent might autonomously handle the repetitive but critical tasks of system maintenance, freeing up human developers for higher-level innovation. When these categories are applied correctly, the deployment becomes a calculated strategy rather than a shot in the dark, ensuring that the right level of autonomy is applied to the right business problem.
In an era where open-source AI tools allow almost anyone to generate software code, the “secret sauce” of a product seems less valuable than it used to be. Why is the speed of execution now considered the only sustainable source of differentiation for modern businesses?
We have reached a point where the intellectual capability to know what to do is a huge asset, but it is no longer enough to stay ahead of the competition. As the barriers to technical execution fall, the only real sustainable source of differentiation from a technology perspective is the sheer speed at which an organization can execute its ideas. If you have a brilliant concept but your internal processes are bogged down by a lack of agility, you will lose to a competitor who can move from conception to a functional product in a fraction of the time. This creates a high-pressure, fast-moving environment where the ability to take an idea and turn it into a reality using agentic tools is the ultimate competitive moat. It is no longer about who has the best “recipe” in their code, but who can cook and serve the meal before the customer moves on to someone else.
The tension between moving fast and maintaining safety is a classic corporate dilemma, especially with a technology as unpredictable as AI. How can a standardized platform model help organizations relax their legal and security bottlenecks without compromising on ethics?
The conventional way of maneuvering through policy involves a heavy presence of lawyers and security experts who often act as a brake on innovation because every project is treated as a unique, high-risk event. However, if you deploy standardized platforms and have a proper governance structure in place, you can actually relax some of those restrictive layers and allow your team to move much faster. Dell is rolling out a principle where trained developers working on trusted, approved platforms are given significantly more latitude because the environment itself provides the necessary safety. This move toward predictability and robustness in models and runtimes works hand in hand with risk mitigation, effectively speeding up the time it takes to get quality services to market. It changes the atmosphere from one of constant “policing” to one of “trusted execution,” where the guardrails are built into the foundation of the technology rather than being an external obstacle.
Moving toward a protocol-based approach for autonomous systems represents a significant shift from overarching high-level policies. What are the specific actions a business must take to build a “deeply trusted” system that operates under these protocols?
A protocol-based approach focuses on the granular level of how probabilistic autonomous systems function, requiring businesses to establish hard technical rules rather than vague policy statements. This starts with issuing a unique digital identity for every agent, ensuring that every action can be traced and accounted for within the system. You must ensure agents work in safe, predictable, and controlled environments where their boundaries are clearly defined and enforced by the infrastructure. By using strong protocols, we ensure that agents perform sufficient reasoning to guide their processes and use supplemental tools that provide a higher degree of predictability. It is about creating a “wrap” of governance—both intrinsic within the agent and extrinsic in the infrastructure—that makes the entire operation feel as secure and reliable as a traditional software system.
Ethical considerations are often complicated by the fact that human values are plural and context-dependent rather than a single monolith. How should developers manage the trade-offs and choices inherent in aligning agentic AI with these evolving values?
We must acknowledge that there is no single, broadly defined set of human values that we can simply “plug in” to a model; values are plural, context-dependent, and constantly evolving. Developers are the ones who shape these values through their choices regarding data sets, objectives, and the inevitable trade-offs they make during the design process. This is why having trusted, educated individuals who manage and work with agents is perhaps the most important facet of agentic AI ethics. It requires both organizational and administrative governance to ensure that the choices made by the AI reflect the specific values of the business and the community it serves. Ultimately, the focus must shift to a more granular, pluralistic understanding of ethics where we don’t look for a “perfect” alignment, but rather a robust and transparent one that respects the complexity of human life.
What is your forecast for the role of agentic AI in the future of enterprise networking and operations?
I believe we are on the cusp of seeing a complete transformation where the “coordination agent” becomes the central nervous system of the enterprise, managing swarms of specialized agents that handle everything from network optimization to cybersecurity. We will move away from humans managing tools to humans managing outcomes, with the AI handling the hundreds of tiny reasoning steps required to reach a goal. The businesses that will thrive are those that have already begun the hard work of setting up their standardized platforms and digital identities today. Within the next few years, an organization without an agentic strategy will find itself moving at a human pace in a world that is increasingly operating at the speed of autonomous reasoning, and that gap will be impossible to close.
