Strategies are being determined. Experimentation is running rampant. Proofs of concept abound. As generative AI quickly gains a foothold across organizations and industries, there’s little consensus yet about how best to determine its impact—and whether C-level executives will reach consensus, themselves. There are clues, however, in how business and technology leaders measure value for traditional artificial intelligence, the larger class of AI investments such as machine learning, deep learning, and conversational AI for which executives have established measurement behaviors and preferences.
Using data from a global survey of 1,600 business and technology leaders across 14 countries conducted in February 2023,1 the Deloitte Center for Integrated Research analyzed how technology leaders and business leaders prioritize the key performance indicators commonly associated with digital investments when assessing the impact of their organizations’ AI capabilities. The results of this assessment proved to be counterintuitive: Interestingly, while business leaders who participated in the survey reported that they’re more focused on AI’s process-related benefits within their organizations, tech leader respondents said they’re more often looking outward—at KPIs associated with sales and customer satisfaction.
According to the survey, technology leaders are 12 percentage points more likely than business leaders to be using the sales of new digital products as a KPI and 7 percentage points more likely to be focused on sales through new digital platforms, for instance. They also use net promoter scores and intangible assets more than business leaders.2
When it comes to all forms of AI, business and tech leaders alike might collectively be missing opportunities to consider innovation measures and long-term value creation, the survey findings suggest. Among those leaders who measure traditional AI, only about 30% use innovation-oriented KPIs like the tech’s effect on an organization’s tolerance for experimentation or intelligent failure, or the number of agile pods or teams.3
Leadership’s alignment on AI success metrics could be less critical during an organization’s experimentation or initial adoption phase, but it could, of course, become increasingly important as the organization works to assess the technology’s current and potential impact, and makes the case for continued investment.
Research and analysis by the Deloitte Center for Integrated Research
Read the full report at www.deloitte.com/insights/measuring-ai.