Analysts at Wedbush Securities said in a Friday (June 26) report in Seeking Alpha that many companies have not developed a way to determine whether the artificial intelligence (AI) tools they deploy are providing a sufficient return on investment.
This is perhaps the most important finding to emerge from discussions at Wedbush Securities’ Disruptive Technology Conference earlier this week, analysts led by Dan Ives said in an investor note on Friday, according to the report.
Analysts learned from executives at the event that companies are investing in AI trials without a framework to measure success, and without such a framework, they are likely to have difficulty justifying their investments, identifying approaches that are working, and building organizational trust in AI-driven decision-making.
“Many executives feel that customers are under increasing pressure from boards of directors and CFOs to prove real benefits from AI, and the inability to answer this question is a significant barrier to additional investment in long-term technology builds,” Ives said, according to the report.
PYMNTS CEO Karen Webster wrote in September that a PYMNTS Intelligence survey found that most business executives have realistic expectations about when they can expect a positive return from their investments in generative AI.
More than eight in 10 executives surveyed said it could take three to 10 years.
“These executives also understand that big ‘T’ transformations typically do not occur on a predictable schedule, nor do they come with the expectation of an immediate or direct ‘million-dollar’ return on investment,” Webster wrote.
Another PYMNTS Intelligence report, “The Enterprise AI Readiness Gap: Corporate Data Reveals the Real Barriers to Scale,” found that when executives were asked whether their organization’s readiness or the capabilities of their AI technology was a greater constraint to AI performance, 71% pointed to the readiness of their organization’s people, processes, or data.
Executives say there are an average of four to five organizational barriers limiting AI performance, with the most common bottlenecks being data quality, budget constraints, and governance processes.
The report states that “piecemeal solutions are unlikely to work as executives raise multiple barriers at the same time.” “Improve data quality, clarify accountability, address talent shortages, and at the same time rethink budgets to make the most of cross-functional AI operating models.”
