AI Spend: The Most Fragmented Line Item In The Enterprise

Jesse Stockall is Chief Architect, Flexera.

​AI has rapidly become one of the most complex areas of enterprise technology spend. Unlike previous waves of IT, AI spend doesn’t reside in a single system or budget. It spans the full technology stack across the end-user applications, underlying models, data platforms and the compute infrastructure that power them. Each layer introduces different consumption models, like licenses, tokens, credits and GPU usage, that are often managed by varying teams with limited coordination.​

This wide dispersion leads to challenges, with Flexera’s recent State of ITAM report finding that only 31% of IT leaders report accurate visibility into their AI software. It’s apparent that many organizations are still unclear on how AI is being adopted, where costs are accumulating and which investments are delivering value. To turn AI from a fragmented expense to a measurable business investment, leaders must first understand where these gaps originate and how to close them.​

AI costs now show up across multiple layers of the technology environment. Employees may use AI-enabled SaaS applications, infrastructure teams may manage cloud resources and data teams may build pipelines to support AI workflows. Each AI-powered action offers a partial perspective into the full scope of activity, but too many organizations are missing visibility into the full AI stack.​

To connect AI activity to business outcomes, organizations need a comprehensive view of tech spend, attributed to users, cost centers and defined outcomes. Organizations will need stronger capabilities to track token usage, cloud infrastructure and model activity as AI adoption scales.

The expeditious pace of AI innovation is changing how organizations pay for technology at a concurrent rate. AI pricing is no longer limited to traditional licenses or straightforward cloud consumption. Enterprises are now navigating a fast-changing mix of tokens, credits, model tiers, usage limits, overages and more inside existing software contracts. Just as challenging, there is little consistency across vendors or services, making it difficult to compare cost or forecast demand.​

As vendors continue to evolve their products and monetization models, enterprises will need to continuously monitor not only what they are spending, but how pricing structures are changing and whether those changes align with their unique business value.​

Recent data from Okta found that two-thirds (67%) of U.S.-based employees use unsanctioned AI tools, and nearly a quarter do so regularly. This reality is creating a growing visibility gap for IT teams around data exposure, usage or spend.

Like shadow SaaS and shadow IT before it, shadow AI is now a problem that can create unpredictable spend, unmanaged vendors, unclear ownership and inconsistent policy enforcement, leading to increased risk. While AI experimentation is not inherently a problem, the challenge for leaders is in ensuring that experimentation becomes visible and measurable through integrated platforms and third-party tools, while still fostering a culture that promotes AI exploration.

To drive change, leaders must focus on closing the visibility gap through a combination of cultural alignment, proactive enablement and improved oversight. Without all three, autonomous AI systems will continue to operate unpredictably, creating risks that undermine trust and reduce ROI.

AI is continuously introducing new challenges for enterprises, but it’s possible for organizations to master this new era of tech thoughtfully. To do so, many leading organizations are addressing AI spend by establishing broad visibility across asset classes and user-level cost attribution, which are foundational elements for accurate ROI calculations.

This means connecting software, SaaS, cloud, infrastructure, data and AI usage into a more unified operating view. This merged approach helps leaders better align technical decisions with financial outcomes, optimizing usage, reducing waste and improving the consistency of AI investments.​

Rather than treating AI as a series of isolated tools or experiments, organizations can manage it as a strategic investment portfolio. The goal is to shift from fragmented tracking to unified, cross-functional decision-making.

Enterprises will only realize the full value of AI by understanding and managing AI spend through connected, cross-stack investment rather than simply adopting more tools. Effectively managing this complexity will be a key determinant moving forward in translating AI adoption into sustained business value.

Forbes Technology Council is an invitation-only community for world-class CIOs, CTOs and technology executives. Do I qualify?