The decision to adopt artificial intelligence is increasingly becoming part of routine business planning, but adoption alone says little about whether the technology is producing meaningful business value. A May 2026 analysis by the U.S. Census Bureau found that business AI use remained between 17% and 20% from December 2025 through May 2026, while 20% to 23% expected to use it within the following six months. These figures point to a practical question for business leaders: what, specifically, do they expect AI to accomplish?
Larger organizations were more likely to report using AI. According to the U.S. Census Bureau report, 37% of businesses with 250 or more employees used AI, compared with fewer than 20% of businesses with four or fewer employees. Yet the fact that a company has adopted AI does not necessarily indicate that it is achieving measurable results from the technology.
Reliability is another factor businesses must consider. The National Institute of Standards and Technology, in its 2024 generative AI risk profile, points to the possibility of systems producing false or inaccurate information with a high degree of confidence. This means that even an answer that appears authoritative may need further review before it is used in making an important business decision.
NIST also recommends evaluating a system’s capabilities and limitations in relation to a clearly defined intended use. For executives weighing an AI investment, this provides a useful place to begin. Identifying the task, understanding the consequences of potential errors, and considering the environment in which the system will be used can help businesses determine what needs to be tested before moving toward wider implementation.
Davina A. Faimon, founder of Davina Faimon Consulting, brings a strategic operations perspective to this discussion. With more than a decade of experience working with AI, she believes that business goals should ultimately guide technology decisions. “Sometimes modern AI is not the best-fit solution,” she observes. Her approach centers on defining the problem clearly enough to identify a solution that is appropriately matched to the organization’s needs.
For Faimon, that process begins with identifying what a business is actually trying to improve. She uses the mathematical idea of an objective function to represent goals such as reducing costs, saving time, or increasing output. In her framework, weighing that objective against available resources and operational limitations can help identify which solutions are realistically achievable. The technology decision, she believes, should follow this assessment rather than come before it.
Understanding a company’s data is another key part of the process, Faimon explains. She sees leadership intuition as a useful starting point, while more detailed analysis can help identify where the organization should concentrate its efforts. In her methodology, she describes “blueprinting” an initial picture of the organization before determining whether the available data can support that understanding.
For Faimon, developing a clearer picture of the underlying problem can make the subsequent technology decision easier. She believes organizations should first establish what they want to improve, whether that involves increasing revenue, managing costs, reducing time, or making better use of existing resources. That objective, in her view, provides the foundation for determining where AI may be capable of delivering meaningful value.
The choice of technology should then emerge from that initial analysis. Traditional machine learning, mathematical methods, and modern AI can each have a place depending on the challenge being addressed. Human judgment, she adds, remains essential for deciding whether a particular approach is suitable and whether the resulting outputs actually support the organization’s intended goals.
“We need a combination of the two. It’s an art and a science, blending human thought with the tools,” she explains. Faimon considers oversight to be an ongoing responsibility, with people reviewing outputs, examining unexpected results, and assessing whether a system continues to operate according to its original purpose. Without that human involvement, she cautions, an output that initially appears useful could eventually contribute to an incorrect business decision.
This philosophy also informs her consulting practice, where she offers strategic leadership support and executive coaching alongside optimization solutions when appropriate. She sees value in combining technical expertise with a clear understanding of what an organization is ultimately seeking to accomplish. Rather than focusing solely on a model’s technical performance, she considers whether its use is contributing to the company’s broader business objectives.
As organizations continue to explore AI, Faimon believes that placing greater emphasis on purpose and problem definition could help businesses make more thoughtful investment and implementation decisions. “Knowing what you want to achieve gives you a starting point. Maintaining and evaluating the results helps you decide what your next pivotal move could be,” she explains. “AI can be the foundational part of the solution, but ultimately the business gets to decide where to take it.”
