Having moved beyond this initial excitement, the revolution now faces a sobering inflection point. According to S&P Global Market Intelligence, the share of businesses scrapping most of their AI initiatives jumped to 42% this year, up from 17% last year.
This new phase has been dubbed “the trough of disillusionment,” where the highs of early adoption confront the realities of implementation.
Steve Daly, Senior Vice President Solutions at New Era Technology, has observed this pattern repeatedly across his clients. He explains that much of the disillusionment stems from a gap between expectation and reality: “The reason this disillusionment sets in is because what companies believed they were going to get just doesn’t come to fruition.”
Many organizations find that expectations don’t match results, revealing the need for more than just enthusiasm to succeed.
Yet amid fierce competition, companies that successfully navigate their AI transition stand to gain a decisive edge. Inertia is therefore not an option.
The good news is that this challenge need not be tackled alone. Strategic integrators like New Era Technology have built systematic approaches to help organizations avoid common pitfalls derailing AI initiatives—turning what has historically been a high-risk gamble into a methodical journey toward measurable business value.
The Infrastructure Challenges That Sink AI Projects
AI project abandonment can be highly specific to a company’s situation. Yet the biggest and most fundamental challenge New Era Technology sees is data. These challenges often remain invisible until organizations attempt to scale beyond proof-of-concept.
While media attention fixates on algorithm capabilities and user interfaces, poor data quality remains AI’s Achilles’ heel, causing 42–85% of projects to fail in 2025. Deloitte’s 2024 survey highlights how 80% of AI and ML projects face difficulties tied to data quality and governance. But what about data causes things to go so wrong? After all, companies’ digital footprints are larger than ever.
When Daly drills down into what’s wrong with enterprise data, he points to a core architectural issue: “The big one is still disconnected systems. Lots of companies have several systems—ERP, CRM, HR—and they’re all disconnected. Yet, they want to create a generative AI tool using all that data.”
This scenario captures the fundamental challenge: most organizations’ data landscapes are not designed for AI consumption. These disconnected systems reflect years of deferred infrastructure decisions that organizations previously managed to work around. But AI implementation exposes these weaknesses in ways traditional applications never did.
This fragmentation isn’t new, deferred decisions around security and data management have long created difficulties within many organizations. But AI has made previously manageable problems critical:
“Many organizations kicked the can down the road when dealing with security and data problems. Now generative AI comes, and everyone wants to do generative AI, yet those problems remain.”
Systems not properly orchestrated before AI won’t function any differently with AI. In fact, silos become even more pronounced.
Beyond these technical challenges, Daly points out that a significant barrier to adoption is organizational understanding. “Really understanding what generative AI does and how it helps is a real stumbling block,” he said, adding that, “there’s so much media pressure it creates challenges for organizations in grasping its capabilities.”
This mismatch between hype and comprehension often leads to unrealistic expectations and insufficient planning.
When these realities inevitably clash with inflated expectations, stakeholder confidence rapidly erodes, creating conditions for abandonment, loss of investment, and a halted AI transformation.
The Intelligent Adoption Framework: An AI-Winning Formula
Having observed these failure patterns, New Era Technology took action, creating a framework to help customers successfully integrate AI into their operations.
The result? The Intelligent Adoption Framework—a systematic approach that addresses root causes of abandonment before they derail initiatives. According to Daly:
“With the Intelligent Adoption Framework, we walk companies through a four-step process focused on AI literacy, process mining, and measuring focus areas for generative AI.”
The framework begins with “Envision,” where New Era Technology facilitates discovery sessions that align technology possibilities with genuine business needs—preventing disconnects between expectations and reality.
The “Align” phase tackles organizational factors by taking companies’ big ideas and “aligning them with the organization a little tighter,” grounding ambitious AI initiatives in practical realities. This includes addressing data infrastructure challenges, security requirements, and compliance frameworks before they become project-killing obstacles.
The “Launch” phase reflects New Era Technology’s philosophy of systematic value demonstration: “We want to get some things out there. Build and put them out so we get learning,” Daly said. This helps customers see AI in action and how it generates tangible business outcomes that maintain stakeholder confidence and provide an empirical foundation for scaling.
The final “Evolve” stage institutionalizes continuous improvement by “doing retrospectives, taking what we’ve learned and either expanding existing solutions or moving into other organizational areas.”
Such a systematic approach has proven especially effective when combined with New Era Technology’s Microsoft partnership, which provides it with enterprise-grade AI tools and infrastructure necessary to successfully execute the framework across diverse contexts. This was demonstrated when New Era worked with a major university medical school on AI.
“We had a client approach us and ask ‘Can you create our own ChatGPT?’” Daly recalled.
“We said, sure, but it’s going to cost millions and take a year. That kind of froze them.”
Instead of pursuing that complex, expensive route, New Era Technology redirected their energy toward an appropriate solution using Microsoft Copilot Studio. This low-code, no-code tool builds generative AI virtual assistants.
Training the company just took a day, and since deploying Copilot Studio, the medical school has rolled out nearly two dozen AI agents across its operations.
This success illustrates how proper guidance and implementation create the company buy-in essential to avoiding abandoned projects and turning AI into a transformational engine.
Strategic Partnership as Competitive Necessity
Enterprises no longer have the choice on whether or not to roll out AI, as Daly points out, “AI is already in your organization, whether you know it or not.”
What they do have a choice on is whether they want to keep pace with the 58% of business going ahead with AI initiatives or the 42% scraping them. With this gap widening daily, those successful companies are only compounding their AI advantages.
Yet organizations going it alone face a brutal reality: every failed implementation burns budget, erodes stakeholder confidence, and hands competitors more time to establish market leadership. The costs extend beyond wasted investment to include missed opportunities that may never return.
Strategic integrators offer a different path. They bring proven frameworks that turn high-risk gambles into systematic business improvements. Rather than learning through expensive mistakes, companies can leverage battle-tested approaches that deliver results while competitors struggle with abandoned pilots.
In an era of rapid technological change and intensifying competitive pressure, strategic partnerships with proven integrators may determine which organizations thrive in the AI future—and which become cautionary tales of missed opportunity.
