Research Reveals AI Gaps in Federal IT Modernization

Yuichiro Chino/Getty Images

Only one-quarter of study participants saw AI as critical or very valuable to their modernization efforts, and one-third saw it as not very/not at all valuable. 

All federal agencies are focusing on IT modernization to keep pace with changing mission requirements and citizen demands but one of the biggest modernization challenges program managers and mission owners face is time spent chasing problems rather than moving the mission forward. Artificial intelligence-based observability and application monitoring can augment agency resources and free up capacity to focus on accelerating mission innovation. Yet despite the clear benefits of this technology, agencies are not embracing AI solutions that can automate manually intensive trouble-shooting and get missions quickly back on point. Some recent research involving 300 IT and procurement professionals in federal civilian and defense agencies provides insight into why that may be the case and where their efforts are being prioritized. 

The Focus is on the Cloud

A fall 2021 study by Market Connections showed that agencies’ most important areas for IT resource investment are cloud computing and application performance monitoring. In fact, four in 10 see cloud computing investment as essential. Following a decade of learning, agencies’ decisioning process for moving to the cloud has matured. Most now apply a deliberative workload migration approach that is driven by mission requirements.

Still, there are a variety of hurdles to overcome. For instance, 48% of study participants saw cost savings as a big cloud benefit, yet 51% said budget is an obstacle. While these opinions seem to be at odds, there will in fact be significant long-term cost efficiencies and security benefits when cloud adoption is done right and once systems normalize. 

Those gains depend on advanced technology, including AI. Cloud migration compounds an already complex IT ecosystem. It impacts access and control over underlying infrastructure, changes how to handle and manage security processes and policies, and complicates identifying and quickly responding to problems.  

Agency cloud strategies will simply not be sustainable for the long term without AI-based “force-multiplier” platforms that provide deeper capabilities to manage the increased complexity. In fact, more than eight in 10 professionals said they are at least somewhat (or more) likely to use an AI platform to help. Such tools will also help agencies meet the Biden administration’s recent Executive Order on Improving the Nation’s Cybersecurity

Monitoring Is a Challenge

Monitoring and visibility were deemed very important by over 60% of government professionals surveyed. While it’s impossible to sufficiently track, monitor and report without modern visibility approaches, on average study participants reported having visibility into only two-thirds of their entire IT environment (the most for network and operating system performance, and the least for virtual machines and cloud instances). That leaves the other third in the dark.

Legacy application performance monitoring (APM) tools are just not sufficient for today’s cloud-powered missions. With hundreds of apps and millions of dependencies, continual, detailed status information is essential to sustaining operations. Agencies are likely using legacy tools that provide some dependency and network mapping, but lack the more advanced root cause analysis and automation capabilities needed for the modern environment. 

As agencies accelerate cloud migrations, it will be necessary to replace those legacy tools with AI-driven observability platforms that can provide the needed insights into system performance and issue resolution. Observability is the next generation of APM, designed to give end-to-end visibility of all applications and components while adding a layer of automation to address today’s complex, dynamic environments. AI-based observability collects, analyses, predicts, finds root causes and extends all the way to the end-user. 

Observability will also help agencies mitigate cloud monitoring limitations like a lack of resources.  For instance, with root cause discovery and remediation, AI can do in a few minutes what would take a team of people hours, weeks or more. Automating lower level or repetitive tasks frees personnel to focus on higher stake tasks, essentially increasing headcount without the cost.

The AI Trust Factor 

Perhaps the biggest issue the study revealed was government professionals’ lack of trust in AI.  Only one-quarter of study participants saw AI as critical or very valuable to their modernization efforts, and one-third saw it as not very/not at all valuable. Many felt that their people, processes and governance were simply not suitable or ready for AI utilization. 

One-third of participants plainly said they just don’t trust it. Without human-readable understanding of AI, its decisions and recommendations are suspect. Bridging that trust will come from AI systems that provide three key elements: 

First, they must be transparent, so professionals know that the AI-based system follows a deterministic and dynamic approach, and they can easily understand and explain the logic behind its decisions. 

That means systems must be explainable. Allowing IT teams to see a complete set of circumstances leading up to any decision the AI engine makes empowers them to trust that future actions under an identical set of circumstances would be the same. This reduces wasted troubleshooting and chasing false-positive alerts, while accelerating results.

Lastly, systems must provide high quality data. Supplying a precise known state of how everything across the technology stack operates means the AI is working with good data, removing the risk of biased, tainted decision-making. When the AI suppresses noise and provides precise answers with root-cause analysis and problem playback, resolution can happen in minutes, often before a mission is impacted.

Failure Is Not an Option

In reality, IT modernization that incorporates AI-level capabilities is integral to keeping agency mission fulfillment at pace with evolving requirements and citizen expectations. With growing cloud complexity, a widely distributed workforce and clever, ruthless adversaries always lurking, this isn’t a choice—it’s a mandate. As current research reveals, government professionals who need to make this happen aren’t there yet. There is urgency for greater collaboration between them and industry who is obligated to make AI solutions transparent, explainable and high quality. The future of the nation depends on it.

Willie Hicks is public sector chief technology officer at Dynatrace.