A survey of 300 senior IT leaders from organizations with 1,000 or more employees in the U.S. and the United Kingdom (UK) finds that many organizations are not yet ready to deploy artificial intelligence (AI) workloads at scale.

Conducted by Global Survey Research on behalf of ControlMonkey, a provider of a platform for automating the management of IT infrastructure at scale, the survey finds 83% of respondents expect the number of AI-driven workloads being deployed will rise 50% in the next 12–24 months. On average, IT teams are forecasting a 50% jump in workload demand, the survey finds.

However, well over half (54%) said they aren’t fully ready for IT automation at AI scale, with top concerns being reliability (43%), skill gaps (39%) and scalability limits (36%). Other concerns include rising cloud costs (27%, overloaded compute and storage capacity (20%), deployment bottlenecks (18%), security and compliance issues (18%) and observability (17%).

A total of 46%, however, also noted they have limited or no bandwidth available to devote to the infrastructure innovation that will be required to deploy AI workloads at scale.

Specific AI workload challenges identified by survey respondents include cost management (37%), lack of real-time infrastructure visibility (36%), and difficulty allocating and scaling resources effectively (32%), security (29%), compliance (29%) and standardizing governance policies (20%).

Nearly half (45%) identified training and visibility as their top need for managing AI workloads, followed by cost controls (21%), governance (20%), and automation (14%).

ControlMonkey CEO Aharon Twizer said the survey makes it clear that a coming wave of AI workloads is only going to exacerbate an existing set of IT management challenges that many IT teams have yet to address. The longer it takes to address these issues, the bigger the problem will actually become, he added.

Unfortunately, the survey found that less than half of respondents said their current IT environments are managed using infrastructure-as-code (IaC) tools and frameworks. While 80% say they have achieved at least a moderate level of automation, only 1% claimed infrastructure management is fully automated.

As a provider of an IT infrastructure automation platform based on open source Terraform IaC tools, ControlMonkey is making a case for modernizing the management of IT environments in advance of the arrival of an onslaught of AI workloads. In the absence of that capability, it will be all but impossible for IT teams to manage an exponential increase in the number of new applications or existing applications that are being updated to add AI capabilities, noted Twizer.

Less clear is the degree to which IT teams might now be prepared to centralize the management of IT infrastructure. Many organizations are now embracing platform engineering as a methodology to achieve that goal, but the overall pace at which those efforts are being made may not be able to keep up with the rate at which additional applications and workloads are about to be deployed. Most IT teams are already clearly falling behind, said Twizer. Without proper guardrails enabled by automation in place, most of them will continue to find themselves in a constant state of firefighting, he added.

Hopefully, challenges involving the management of workloads at scale will be addressed before there is an actual crisis, but if history is any guide, it might not be until there is an actual crisis before many IT teams will find a way to rise to the challenge.