Measuring AI Use Becomes a Business Requirement
As enterprises increasingly integrate artificial intelligence (AI) tools into their daily operations, the need for effective measurement and governance has become a pressing business requirement. Organizations are utilizing a myriad of AI applications for tasks ranging from code generation and analytics to customer support and internal research. However, oversight of these tools remains inconsistent across various roles, functions, and industries.
The Visibility Gap
A recent survey highlights a significant discrepancy in perceptions of AI visibility within organizations. Executives often express confidence in their understanding of AI activities, while directors and managers who are closer to daily operations report a contrasting reality. This results in a 16-point gap in confidence regarding AI visibility, a trend that is evident across different industries and company sizes.
Shadow AI Usage
Another contributing factor to this disconnect is shadow AI usage. More than one-fifth of leaders identify the use of personal or unsanctioned AI tools by employees as a barrier to success. Interestingly, most leaders reporting this barrier also express high confidence in their visibility of AI activities. While tool procurement can provide insights into licenses acquired, it offers limited visibility into daily usage patterns.
Executive vs. Operational Perspectives
Russ Fradin, CEO of a leading research firm, pointed out, “The C-suite believes AI is visible, valuable, and under control, while adoption is racing ahead of measurement, and governance is inconsistent.” This misalignment indicates that until enterprises can organize their efforts around real-time data, AI may pose both a strategic asset and a liability.
The Tool Landscape
Most enterprises today rely on multiple AI products to enhance their operations. Organizations that report stronger returns typically utilize an average of 2.7 tools, significantly more than the 1.1 tools used by lower-performing peers. These specialized tools cater to various workflows such as software development, automation, and content generation. However, this diversification can lead to redundancy and budget waste as overlapping tools are often seen as unnecessary expenses.
Inventory and Governance Challenges
Despite the growing adoption of AI tools, only 38 percent of organizations maintain a comprehensive inventory of AI applications in use. This lack of visibility complicates governance, budgeting, and risk management, particularly as regulations such as ISO 42001 necessitate continuous awareness of deployed systems.
Return on Investment by Sector
Return on investment (ROI) from AI tools varies significantly across sectors. Industries such as retail, software, manufacturing, and telecommunications report a high likelihood of realizing ROI within six months. In contrast, sectors like hospitality and healthcare tend to report lower expectations, largely due to the structural complexities behind their workflows.
Job Function Insights
Results also differ by job function. IT teams, for instance, report the strongest outcomes and the highest confidence in both visibility and ROI. They leverage AI to generate code, automate infrastructure, and accelerate delivery, yielding measurable results in deployment frequency and system uptime. Conversely, customer support roles exhibit lower confidence and ROI, despite heavy investments in AI tools like chatbots.
Understanding Productivity Gains
Most workers report only modest time savings from AI, with over 85 percent indicating savings of less than 10 hours per month. Only a small group of power users—approximately six percent of the workforce—claim to save more than 20 hours per month, engaging across multiple tools for advanced capabilities. Training is closely correlated with proficiency, where organizations with formal AI training programs report higher skill levels and productivity gains.
Structural Measurement Issues
A lack of responsibility for AI measurement is evident, with 30 percent of respondents noting gaps in accountability. Fragmented ownership across teams exacerbates the situation, although governance policies are present in most organizations. About 69 percent have AI risk and compliance policies; however, many lack visibility into adoption rates and risk exposure.
Conclusion
The current state of AI utilization within enterprises underscores the need for comprehensive measurement and governance strategies. As organizations continue to adopt multiple AI tools, understanding their effectiveness and aligning them with business outcomes will be critical for leveraging AI as a strategic asset.
