By PYMNTS | June 26, 2026
As the initial frenzy of generative AI adoption settles into a phase of pragmatic implementation, a sobering reality is beginning to take hold in corporate boardrooms across the globe. Despite the massive capital expenditure funneled into artificial intelligence pilots over the past two years, a significant majority of enterprises have yet to develop a robust, quantifiable framework to determine whether these investments are actually delivering a return on investment (ROI).
This findings, highlighted by analysts at Wedbush Securities following their recent Disruptive Technology Conference, suggests that the "gold rush" era of AI experimentation is reaching a critical inflection point. As CFOs and boards of directors demand greater fiscal accountability, the lack of clear performance benchmarks is creating a widening gap between technological promise and operational reality.
The Core Problem: Measuring the Immeasurable
The primary takeaway from the Wedbush Disruptive Technology Conference, held earlier this week, was not a lack of AI capability, but a systemic lack of measurement maturity. Led by renowned analyst Dan Ives, the Wedbush team reported that many enterprise executives are currently navigating their AI strategies without a compass.
The fundamental issue is twofold: many organizations launched AI pilots as reactive measures to competitive pressure, rather than strategic initiatives tied to specific business outcomes. Consequently, these pilots often lack the key performance indicators (KPIs) necessary to justify their transition from experimental "sandboxes" to enterprise-wide production environments. Without a standardized framework to gauge success, businesses are finding it increasingly difficult to discern which AI applications are driving efficiency and which are simply inflating operational costs.
"Many executives noted that customers are feeling increased pressure from their boards and CFOs to demonstrate actual returns from AI," Ives stated in a Friday investor note. "The inability to answer this question presents a real barrier to additional investments in long-term technological buildouts."
Chronology of the AI Investment Cycle
To understand why enterprises find themselves in this current state of uncertainty, it is necessary to examine the evolution of AI adoption since the late 2022 explosion of generative tools.
- Phase 1: The Reactive Sprint (Q4 2022 – Q2 2023): Following the public release of advanced large language models, enterprises scrambled to implement AI tools to avoid being left behind. During this period, "speed to market" was the primary metric of success, and many firms bypassed traditional rigorous vetting processes in favor of rapid prototyping.
- Phase 2: The Pilot Proliferation (Q3 2023 – Q4 2024): Organizations moved into a period of heavy experimentation. Numerous pilots were launched across departments, from customer service chatbots to automated software coding assistants. However, because these were often decentralized, there was little consistency in how their value was recorded.
- Phase 3: The Reality Check (2025 – Present): As budgets for 2026 and beyond are being scrutinized, the "AI-first" mandates are meeting the "bottom-line-first" reality. Enterprises are now being forced to reconcile their spending with actual impact. The current discourse has shifted from "how can we use AI?" to "how is this AI tool impacting our EBITDA?"
The Long-Term Horizon: A Mismatched Expectation
One of the most persistent issues in the current AI landscape is the misalignment between executive timelines and the reality of organizational transformation. While boards often push for quarterly updates on AI profitability, the nature of such "Big-T" technological transformation rarely adheres to such short-term schedules.
Data from PYMNTS Intelligence provides a crucial counterpoint to the current panic. In research conducted last September, it was revealed that most enterprise executives actually possess a realistic—if sobering—understanding of the payback period for generative AI. More than 80% of surveyed executives acknowledged that realizing a positive return on investment could take anywhere from three to ten years.
This suggests that the frustration currently being expressed by boards may stem from a communication breakdown between those managing the technology and those overseeing the capital. As CEO Karen Webster noted, "These enterprise executives also understand that big-T transformation doesn’t usually happen on a predictable timetable, nor with the expectation of an immediate or direct payback in the millions."
Organizational Barriers: It’s Not the Tech, It’s the People
While analysts often focus on the performance of the AI models themselves—such as latency, accuracy, and parameter size—the actual bottleneck to AI performance is far more human-centric. The report, "The Enterprise AI Readiness Gap: What Company Data Reveals About the Real Barrier to Scale," found that 71% of executives identified organizational readiness—specifically people, processes, and data—as the primary constraint on AI performance, rather than the capabilities of the technology itself.
Key Bottlenecks identified by enterprises:
- Data Quality: AI models are only as effective as the data they are trained on. Siloed, fragmented, and unstructured data remains the single largest technical hurdle for enterprises attempting to scale AI.
- Governance Processes: Many firms lack the internal policies to handle AI ethics, data privacy, and intellectual property concerns, leading to stalled projects and legal risk.
- Budget Limitations: With capital being diverted toward AI, other essential IT infrastructure projects are often neglected, leading to a "hollowing out" effect where the AI tools lack the backend support to function effectively.
- Talent Gaps: There is a persistent shortage of personnel who understand both the nuances of AI engineering and the specific operational requirements of the enterprise.
Implications: The Move Toward "Cross-Functional" Maturity
The conclusion drawn from the current state of the market is that piecemeal problem-solving is no longer a viable strategy. Enterprises that attempt to solve for "data quality" in one silo while ignoring "talent gaps" in another will continue to see their AI initiatives languish in the pilot phase.
To achieve long-term success, organizations must adopt a cross-functional operating model. This involves:
- Parallel Development: Addressing data infrastructure, talent acquisition, and governance simultaneously rather than sequentially.
- Clarified Responsibility: Defining exactly which department owns the AI outcome. When AI is "everyone’s project," it is often effectively "no one’s project."
- Rethinking Budgets: Moving away from experimental, ad-hoc funding toward long-term capital allocation that mirrors other core infrastructure investments, such as ERP or cloud migrations.
Official Responses and Industry Outlook
The consensus among industry analysts is that the coming 12 to 18 months will be a "shakeout" period. Enterprises that cannot clearly articulate the value proposition of their AI investments will likely see those projects defunded. Conversely, those that successfully implement rigorous, outcome-based measurement frameworks will be better positioned to scale.
"The winners in the next phase of the AI cycle will not necessarily be the ones with the most advanced algorithms, but the ones with the most advanced management practices," said a spokesperson for a leading technology consultancy. "You cannot manage what you cannot measure, and currently, the majority of the market is in the dark."
For stakeholders, the message is clear: the era of "AI experimentation for the sake of innovation" is ending. The era of "AI implementation for the sake of efficiency" has begun. Boards and CFOs will continue to tighten the leash on spending until the industry can move past the current phase of ambiguity and provide concrete proof of value.
Whether this transition leads to a cooling off of the AI market or a more sustainable, value-driven surge in investment depends entirely on how quickly enterprises can bridge the gap between their technical potential and their organizational readiness. As the data suggests, the challenge is not in the silicon, but in the strategy.
