IC Realtime CTO Andrew Nassar contributed expert commentary to CIO.com on how organizations are setting boundaries, measuring outcomes, and “iceboxing” AI projects that don’t prove value fast enough.
As AI adoption accelerates faster than most IT budgets can keep pace, many technology leaders are being pushed to fund new initiatives without meaningful budget expansion. The result is often a zero-sum exercise: delaying some work, consolidating tools, and reprioritizing projects to create room for AI experimentation and deployment.
In the CIO.com feature, the central tension is framed as a trade-off between short-term stability and longer-term capabilities. Leaders are balancing executive pressure to “do AI” with the operational reality that infrastructure refreshes, platform work, and non-AI roadmaps still carry risk if deferred too long.
Andrew Nassar’s commentary focuses on what disciplined AI demand management can look like when requests surge from outside IT. He describes a shift from ad hoc purchasing to tighter evaluation standards—prioritizing tools that demonstrate near-term operational efficiencies over highly experimental efforts.
That approach is translated into measurable planning: defining goals up front and setting checkpoints along the way. Nassar notes that when a project doesn’t perform against expectations, it may be paused rather than pushed forward indefinitely, with a typical proving window described as roughly a quarter.
One example highlighted is a proposed customer support initiative that included an autonomous sales agent platform and a support chatbot. During a pilot, feedback indicated customers weren’t finding support articles and call volume increased, while the projected cost to deploy was described as substantial—factors that contributed to shelving the effort.
Nassar’s commentary also emphasizes the operational overhead that can be overlooked in AI proposals. Beyond licensing, systems may require ongoing maintenance, configuration, and continuous tuning—particularly for customer-facing experiences where tone and reliability can affect brand risk.
Within the broader piece’s theme, the takeaway is that budget strategy is only part of the story. Governance, risk, and execution capacity can be equally decisive in determining which AI projects move forward, which get delayed, and which are paused until the foundation is stronger.