Businesses are facing increasing pressure to explore automation across customer service, lead qualification, scheduling, and core operations as AI capabilities become more widely available. The appeal is clear. Repetitive tasks can be completed faster, information can be processed at scale, and teams may have more time to focus on higher-value responsibilities.
However, the broader conversation around AI suggests that automation by itself may not be the right goal. According to the Boston Consulting Group’s 2026 report, 50% to 55% of U.S. jobs could be reshaped by AI over the next two to three years. The report also estimates that 10% to 15% of jobs could potentially be eliminated over a longer period. These figures point to a business environment where technology decisions increasingly involve work design, skills, accountability, and human judgment, as well as efficiency.
Matthew J. Fitzpatrick, owner of Fitz Enterprises, a technology advisory firm specializing in IT, cybersecurity, compliance, and AI strategy, believes this distinction deserves greater attention. His perspective suggests that businesses may gain more by identifying where automation creates real value and where human involvement remains necessary.
This philosophy begins with two areas Fitzpatrick believes deserve particular consideration. The first is expanding automation into activities that depend heavily on trust, relationships, empathy, or nuanced decision-making. Customer interactions, executive judgment, and complex problem-solving often involve elements that are difficult to translate into a workflow. AI can assist with preparation, information gathering, and routine tasks while people retain responsibility for important decisions.
The second area involves the policies and safeguards that guide AI use. As employees experiment with tools through personal accounts and mobile devices, proprietary or confidential information may enter systems without suitable organizational controls. For Fitzpatrick, AI governance is therefore a critical part of the technology conversation. Security policies, access restrictions, and responsible-use procedures can provide a framework for experimentation while helping organizations understand where sensitive information is being shared.
The quality of the underlying process is just as important. “Poor inputs lead to poor results, and automation just scales the problem,” Fitzpatrick explains. “If a workflow is inefficient, automating it usually just lets that inefficiency run wild at top speed. You really need to step back and evaluate your processes before jumping into buying tools.”
His approach has also been influenced by personal experience. A motorcycle accident and the recovery that followed became part of a broader lesson about transformation, consistency, and progress. The experience reinforced the value of taking practical steps consistently instead of waiting for the perfect moment. “A phrase shared with me in a business setting, ‘Done is better than perfect,’ became another useful reminder that excessive pursuit of perfection can delay meaningful progress,” Fitzpatrick shares.
He brings that mindset to technology through a clear sequence: understand the desired business outcome, examine the workflow, identify the constraints, and then determine where technology can help. The central question is how automation can improve business performance while preserving the human contribution that makes a difference.
A practical decision-making model can emerge from this approach. Highly repetitive, high-volume activities may be suitable for AI when the associated risks and consequences can be properly managed. A hybrid approach may be more effective when AI organizes information or accelerates analysis while people maintain control over decisions and execution. High-trust interactions may still require significant human involvement, particularly when empathy, context, and relationship-building matter.
This strategy-led philosophy can also influence how organizations introduce AI. Fitzpatrick describes a tailored roadmap that starts by assessing workflows, operational bottlenecks, repetitive tasks, and potential opportunities for improvement. Businesses can then prioritize areas that are likely to deliver meaningful returns before gradually expanding into other operational functions, with stronger security governance and regular reassessment in place.
The broader research helps explain why a measured approach matters. Boston Consulting Group’s analysis suggests that AI’s impact will vary considerably across occupations. Some roles may become more AI-enabled, others may be redesigned, and some could face greater substitution. The report estimates that 23% of current jobs fall into an “enabled” category, where AI may become part of everyday work while core responsibilities remain largely unchanged. Another 34% fall into a limited-exposure category in its model, where the potential for automation appears more restricted.
For business leaders, these distinctions offer a useful framework for planning technology investments. AI adoption can become less about introducing tools for their own sake and more about deciding how technology, process design, security, and human expertise can work together effectively.
Fitzpatrick puts it simply: “My job isn’t to sell you more technology. It’s to help you make the right technology decisions for your business.” In an increasingly AI-driven business environment, this perspective reflects a form of technology leadership that goes beyond selecting tools. It requires understanding the business first, putting appropriate safeguards in place, and recognizing where human judgment continues to provide value.
AI will likely keep expanding the range of tasks technology can support. Organizations that navigate this expansion thoughtfully may be those that give equal attention to strategy, governance, process quality, and the people responsible for delivering results. In that environment, embracing AI does not have to come at the expense of keeping human judgment at the center.
Email: [email protected]
