AI won’t transform your business if you’re still running it the same way

Source Summary

Many organizations have seen real gains in productivity and automation from experimenting with AI. But only 34% are using AI to deeply transform their businesses, according to Deloitte’s 2026 State of Generative AI in the Enterprise report. Meanwhile, 37% are using the technology at a surface level with little or no change to underlying business processes. That may explain why so many organizations are still waiting for the transformative ROIs they expected. We’ve seen this before. During the process reengineering movement of the late 1980s and early 1990s, and again during the dot-com era , organizations invested heavily in new technologies and new ways of working. Many failed, not because the technology was flawed, but because they were unwilling to rethink how the business itself operated. A textile manufacturer learned that lesson the hard way more than 30 years ago. The company implemented software designed to support a fundamentally different way of doing business but insisted on preserving decades-old workflows and management practices. The technology was expected to conform to the business, rather than the business adapting to the technology. The implementation failed. Many companies are at risk of making the same mistake with AI because they rely on a bottom-up approach, where employees find ways to use the technology to solve the problem of the day: writing emails, summarizing meetings and accelerating familiar workflows. Top-down transformation starts with a harder question: if AI had existed when we built this company, would we have designed the business this way? The organizations seeing transformative returns are the ones rethinking how their business operates from the ground up, not just streamlining existing workflows. Four reasons AI transformation stalls Organizations often assume that providing access to AI tools will naturally lead to transformation. The reality is that people, incentives and mindset are what determine success. 1. Organizations reward the wrong behaviors One of the fastest ways to derail transformation is to reward people for preserving the status quo. The textile manufacturer encountered this problem when it redesigned its manufacturing ordering system. Leadership wanted greater visibility across the production process and a more responsive, just-in-time operating model. But shift managers were still compensated based on how many pounds moved through their individual work centers each day. Their incentives rewarded maximizing output within their own area instead of supporting the broader changes leadership wanted to implement. The lesson applies directly to AI transformation. Organizations often talk about reinventing workflows while continuing to evaluate employees using metrics designed for a pre-AI world. People optimize for how they’re measured. If compensation, accountability and recognition remain tied to legacy processes, employees will naturally protect those processes. Transformation requires aligning incentives with the future state of the business. 2. Communication breaks down in the middle Executives may have a clear vision for transformation, but that vision often weakens as it moves through the organization. At the textile manufacturer, senior leadership understood the goal of becoming a just-in-time manufacturer. The technology team understood it because they were involved in the implementation. Middle management, however, never fully embraced the vision. The result was that executives talked about doing things differently while managers continued reinforcing existing behaviors and employees received conflicting signals about what success looked like. Many AI initiatives today face the same challenge. Leaders announce ambitious transformation goals, but managers continue operating under assumptions built around the previous way of working. AI transformation requires both top-down direction and bottom-up execution. The middle layers of the organization serve as the connective tissue between the two. Without that connection, transformation efforts quickly become technology projects rather than business initiatives. 3. Training focuses on tools instead of transformation Many organizations approach AI training primarily as a technology exercise. Employees gain access to a new tool, and training focuses on how to write prompts, use copilots or navigate the new application. Those skills are important, but they are only part of the equation. At the textile manufacturer, technology teams needed a deeper understanding of how the production floor actually operated. At the same time, business leaders needed a better understanding of what the technology could enable. Neither side could successfully redesign the process on its own. A similar dynamic exists with AI. Technology teams need business context, and business teams need technology context. Organizations that can bring those perspectives together through cross-functional teams focused on solving business problems rather than technology implementation are the ones making the most progress. 4. People need permission to work differently One of the least discussed barriers to AI adoption is psychological. Many people still associate their value with effort; they take pride in the time, expertise and work required to complete a task. When AI reduces that effort, some employees become uncomfortable acknowledging its role. For some, admitting AI helped feels like diminishing their contribution, which is why leadership visibility matters. Employees need to see leaders openly using AI, sharing examples and discussing how it is helping them work differently. They need to hear that the goal is not simply working faster but applying judgment, creativity and expertise in higher-value ways. AI transformation is ultimately a mindset shift. People need permission to redefine what productive work looks like. Transformation requires more than upskilling Much of the conversation around AI focuses on upskilling. While new skills are important, they are not the primary obstacle to transformation. The bigger challenge is creating a workforce that wants to participate in it. Some employees will embrace experimentation, seek new opportunities and help shape the future of the business. Others will continue looking for ways to preserve the processes that made them successful in the past. Leaders need to recognize the difference and create opportunities for the right people to rise to the occasion. Employees with a fixed mindset will resist change regardless of the tools available. The organizations that succeed will communicate not just what they’re trying to accomplish, but why. Many employees assume AI initiatives are purely about efficiency. The message from leadership needs to be different: we are rebuilding how this business operates, and you are part of that. Increasingly, everyone has access to the same AI tools. Two organizations can deploy the same technology and achieve dramatically different outcomes depending on how they align incentives, communicate expectations and rethink long-standing business processes. Companies that treat AI as a way to make existing work more efficient will continue to see incremental gains, while those willing to question whether that work should be done the same way at all will discover entirely new ways to operate. This article is published as part of the Foundry Expert Contributor Network. Want to join?