Generative AI (GenAI) is here to stay, with organizations worldwide relishing the technologies’ capabilities. Already, 72% of organizations report currently using GenAI either extensively or sparingly and another 26% are experimenting with the technology. However, this new stage of GenAI adoption is still in the early days.
According to McKinsey, only 1% of company executives describe their GenAI rollouts as “mature,” meaning the technology is fully integrated into workflows and drives substantial business outcomes. Closing this maturity gap requires continual course correction, often coming down to deployment roadblocks, such as significant expenses, distrust of unproven technologies, and regulatory risks. If these challenges sound familiar, they should – when IT teams first jumped to embrace the cloud as the next big thing, many of the same barriers surfaced.
The two waves of new technology fervor do differ in some ways. While cloud computing was implemented in more mission-critical systems early on, GenAI is being adopted more rapidly in pilot stages and for use cases primarily devoted to efficiency and productivity gains. Yet, the learning curve is similar: they both push organizations to think and work differently.
By reflecting on the experiences of their cloud computing predecessors, today’s GenAI hopefuls can position themselves for a better-informed future.
Managing Cost, Risk, and Change: Learning from Cloud Missteps
Turning back the clock to when cloud technology started gaining traction, many organizations underestimated the complexity of migration and overestimated short-term cost savings. As a result, most of those same organizations fell victim to three main pitfalls: poor cost management, security misconfigurations, and the natural resistance that comes with cultural and organizational changes.
The cloud era taught us that simply “lifting and shifting” workloads – moving them to the cloud without modernization – often failed to deliver value. Similarly, GenAI initiatives frequently stall when organizations attempt to plug legacy, unstructured, or poorly documented data into powerful new models without updating data foundation. In fact, GenAI projects can deliver underwhelming results or even reinforce existing inefficiencies. The lesson: technology alone cannot overcome foundational weaknesses.
Just as cloud technology exposed gaps in governance, skills, and long-term strategy, so has GenAI. Should employees adopt GenAI tools without oversight or use the technology outside the bounds of acceptable use policy, the risks of shadow IT may reappear, along with the difficulties of securing GenAI pipelines and ensuring compliance at scale. These parallels will continue to surface as GenAI moves from experimentation to widespread enterprise integration, requiring the same robust cybersecurity frameworks, incident response plans, and governance structures found within the cloud.
Beyond risk management, unmanaged cost-sprawl is a longstanding issue in tech. The cloud is no exception and as businesses continue to integrate GenAI into their workflows, they face a similar escalation in expenses.
A growing number of organizations trying to improve their cost management strategy are turning to FinOps as a solution. Leveraging timely, data-driven insights to help improve forecasting and encourage cross-functional accountability and collaboration, a comprehensive FinOps infrastructure has proven invaluable for curbing overspending and maximizing business value. FinOps principles are not limited to cloud cost management alone, offering a viable option for GenAI spending as well.
Putting Cloud Lessons into GenAI Practice
By the end of this year, Gartner predicts at least 30% of GenAI projects will be abandoned after proof of concept. When hype outpaces reality, hidden patterns behind GenAI project failures – like unprepared data, unclear business ownership, or unnecessary complexity often go unnoticed in the rush to adopt new technology. Recognizing and addressing these signals early can mean the difference between GenAI success and another abandoned project. Leaders who actively watch for these warning signs, rather than shortcutting the process, set their teams up for long-term success.
Once adoption is approved, companies should put an emphasis on small GenAI pilot projects to test and ensure real-world value instead of jumping to immediate enterprise-wide scaling. It’s critical that companies begin with only a few clearly defined, high impact use cases with clear ROI goals mapped back to real business needs.
This ensures early wins, builds internal confidence, and avoids wasting time and resources on generic experimentation. By anchoring GenAI adoption to a tangible outcome – like automating customer support summaries or accelerating code reviews – organizations can demonstrate value quickly, refine their approach, and scale more strategically. It also helps align technical efforts with business goals, which is where many GenAI pilots currently fall short.
From there, establishing strong checks and balances, ongoing monitoring, and clearly defined governance policies is the next critical step for responsible use and compliance. Engaging with external experts can be a great first step in navigating today’s complex and ever evolving regulatory landscape. By investing in the right tools and infrastructure early in the GenAI implementation process, along with continuous training, organizations can set the foundation for sustainable success.
Striking the Right Balance with GenAI Innovation
By applying cloud-era lessons with discipline and foresight, organizations can avoid costly missteps and unlock GenAI’s full potential – securely, sustainably, and at scale.
GenAI is set to remain a powerful force, with 70% of CEOs reporting that they expect the technology to impact their business models over the next three years. A number that increases to 89% among those already using the technology. Clearly GenAI’s transformative potential is proving valuable to executive decision makers, but sustainable, large-scale impact still depends on addressing trust, governance, and integration barriers.