The daily routine of an IT professional remains largely reactive. A lot of time and budget is spent responding to tickets, limiting the time that we spend creating systems that drive growth and improve business value. I’ve long said, “the tyranny of the urgent leaves little time for the truly important.”
Multiple factors contribute to this reactive nature and diminish IT teams’ ability to contribute to operational resilience and maturity. According to data from a recent SolarWinds report, more than half of IT professionals (53%) say inefficient workflows slow their ability to respond to issues. More than 1 in 3 (36%) professionals say understaffing is a key challenge.
The fast-paced growth of IT environments, never-ending to-do lists, and the implementation of GenAI (shadow or otherwise) contribute to further difficulties for IT professionals. What today’s IT professionals need is a helping hand—an extra team member or members—that can streamline repetitive workflows, increase efficiency, and give IT pros time back to focus on strategic initiatives. In other words, it’s time for IT leaders to add a new member to their team: the AI agent.
Understanding the AI agent
AI agents, also called “agentic AI,” are AI systems that autonomously make decisions and/or execute tasks. They are the next evolution in how organizations automate and further maximize their IT technology and data estate. There are several types of agents which can range from low-level functional agents, which only execute simple commands, to multi-agent systems, in which multiple agents work together using many data sources and inputs to achieve a higher level of functionality.
Note that while AI agents might be seen as additional coworkers, the goal is not to replace talent. Rather, the goal of agentic AI is to make us as productive as possible by eliminating as many manual, trivial tasks from our workloads as possible. In other words, the helping hand of our AI agents holds the promise of finally defeating the tyranny of the urgent, enabling us to focus on the work that truly adds value to the business.
What does this look like for IT pros?
Agentic AI can mitigate many of the aforementioned issues. For example, with the right agentic AI solution, IT pros can streamline incident response. Instead of having team members manually review every trouble ticket, our AI agent can receive the ticket, recognize the error message and priority level in the ticket details, then initiate remediation. This autonomy of action dramatically improves important metrics, such as MTTx (mean time to detect, acknowledge, and resolve), giving valuable time back to your IT teams.
In the same way, AI agents can also help IT teams provide excellent post-incident reports. If an outage occurs, an agentic system summarizes the events and metrics of the outage, research knowledgebases for remediation details, and suggest potential root causes. Post-incident reports of this kind take a lot of time from an experienced help desk analyst to create when done manually. Agentic AI instead gives us a starting point for incident response and leaves us with time to work on improvements to the IT environment or time for strategic planning to support our long-term goals.
Responsible AI Agent Implementation
For agentic AI, or any AI, to work effectively, it should integrate with the information technology we use daily, from tooling, to data, to applications running throughout our entire IT environment. In this way, AI agents can automate so many of the tasks we do as IT pros. However, this necessary level of access is why it’s important to responsibly implement agentic AI before IT teams start to use it.
In previous years, the IT industry adopted DevOps practices to improve speed and productivity of their Dev teams. In recent years, SolarWinds has led the way to new “Secure by Design” practices in DevOps to better protect systems from threats, both external and internal. Our history teaches us that the most important information technologies require putting security and good governance as one of the first steps of our design processes. In the same way, the evolving age of agentic AI also is important for IT teams (and organizations as a whole) to lean on the “AI by Design” framework to set the groundwork for effective agentic AI. Think of it as laying the foundation for a new role before hiring your new agentic coworker(s).
Just as SolarWinds introduced the “Secure by Design” concept to DevOps, here are our five bedrock principles that make up successful AI by Design:
- Privacy and Security — The Privacy and Security principles dictate the rules we apply to personal and corporate data collection, storage, use, and security. This principle informs practices such as privileged access management, multi-factor authentication, and role-based access control. IT teams should also engage in anonymization and pseudonymization techniques to help strengthen privacy and replace identifiable information.
- Accountability and Fairness — Despite the impressive capabilities of agentic AI, AI systems are still known to perpetuate biases that may show up in the data. This could potentially lead to discriminatory outcomes, especially when the AI agent is allowed to recommend a specific course of action or to make decisions by itself. This is one of the reasons it’s important to keep a human in the loop who analyzes and regulates agentic AI outcomes. IT teams should also maintain feedback mechanisms that share when an AI agent has been at the center of a negative user experience or somehow enabled a poor outcome.
- Transparency and Trust — While AI agents work in the background to remove or automate manual tasks, their workflows must maintain a level of transparency that allows IT teams to explain what each agent is doing, enabling them to figure out the root cause when things go wrong. This is known as “explainable AI” and, while it might seem obvious, many commercial AI agents do not currently provide this feature. Only by using explainable AI agents can we maintain trust in each agentic AI system we deploy, while allowing for intervention and correction, if necessary.
- Simplicity and Accessibility — The use of agentic AI should remove as much complexity as possible. If the goal of AI agents is to make the workload lighter for your IT team, it must be simple to use. For example, try to leverage agentic tools or platforms that allow for natural language prompts. A question as simple as “How many tickets have been submitted in the last 24 hours,” should give as detailed a response as a more complicated prompt. Simplicity and accessibility allow IT teams to gain the most benefit from what agentic AI has to offer.
- Autonomy Boundaries and Safety — When implementing AI agents, we must also define and enforce the limits of our agentic systems. Just as many cloud services can deliver unexpectedly high costs (“cloud sticker shock”) when they are unregulated, AI agents can also deliver unexpected results when they are unbounded. By setting strict capability constraints and validating them continuously, we can ensure agents act reliably without slowing the service delivery.
Stepping into the Future of IT Work
Agentic AI is already having positive effects for many organizations. A recent study from PwC of 300 executives found that, of those who’d adopted AI agents, 66% said they were “delivering measurable value through increased productivity.” As IT environments grow and organizations depend more on digital tooling, IT professionals will need more help to keep productivity high and workflows agile, all while avoiding dangerous burnout.
AI agents will soon become a must-have for today’s IT teams. When agentic AI is implemented strategically and responsibly, organizations can create a pathway for their IT teams to continually add value. I look forward to the day when we can all spend less time on those unexpected and urgent tasks and can spend more time on what’s important to our organization.