If they’re going to compete in today’s tech-dominated world, businesses know that they have to invest big in artificial intelligence. But are other key IT expenditures going begging as a result?

There’s no question that AI is taking up a commanding share of current technology spend, affecting virtually every function of the organization. Yet a growing number of tech leaders are finding themselves having to “triage” IT budgets for the coming year because of it, says Jason Rosenfeld, chief growth and alliance officer with NewRocket, a specialist in AI implementation utilizing the ServiceNow platform and Claude, Anthropic’s agentic AI assistant.

The spending priority for many companies now, Rosenfeld says, are AI-driven cybersecurity systems, as well as the deployment of tokens, the data units that allow AI models to read, process and write natural language. Meanwhile, they’re deferring or reducing investment in other technology initiatives.

“We’re seeing companies continuing to shift dollars from investments in what were enterprise platforms to AI,” he says, citing software-as-a-service (SaaS) as one category of applications that’s especially feeling the drain.

To be sure, organizations are dazzled by the huge advances in capability that AI models have chalked up in recent years, and are spending accordingly. But they’re also beginning to question the efficacy of certain legacy technologies.

“They’re tired of investing in SaaS platforms that aren’t giving them the actual business outcomes they were expecting,” says Rosenfeld. And while no company is contemplating a wholesale ejection of those core hosted applications from its tech stack — “[SaaS] is too embedded in the day-to-day” — they are exploring ways in which AI agents can take on many of the tasks that are currently being performed under the SaaS model.

In addition to cybersecurity, AI spend is being directed to basic operating expenses and tasks geared toward actual revenue generation. NewRocket works with many companies held by private equity, which tends to push for expense reduction and is keen on seeing AI agents take on some of the work previously farmed out to expensive third-party contractors. “It’s about creating value for those companies,” says Rosenfeld.

That said, there is a certain “jump on the bandwagon” feel to the current AI spending spree, and many companies aren’t happy with their results to date. The number of cases in which AI has been successfully implemented deep within organizations “is very low right now,” Rosenfeld acknowledges. “We’re not seeing, to a degree, what most C-level executives and board members had envisioned.”

Part of the problem lies in the failure to establish a strong business case for AI up front. Rosenfeld says organizations need to step back and decide just what AI is going to automate, and the measurable business value they can expect to derive from it. Otherwise, they could end up slapping AI on relatively menial tasks, driving up costs far in excess of their anticipated AI budgets.

Companies looking to redirect dollars from traditional IT spend to AI without thinking about the context in which the cutting-edge technology will be deployed are making “a big mistake,” he says. “If your data and workflows are not in good health, then what are you automating? You’re automating crap.”

None of which portends a slowdown in corporate AI spending anytime soon. On the contrary, Rosenfeld expects to see an upward trend in 2027 IT budgets, especially as applied to cybersecurity and corporate governance. And traditional applications, such as those delivered through SaaS, could well see their role in organizations continue to erode.

“Pulling back on [AI] would, one, have a very negative effect on people and, two, from an innovation respect get a ton of pushback,” Rosenfeld says. “What we’ll see is more controlled spend. It’s not going to be about just spending — it’s about tying that spend to value.”