Why Can’t 82% of Executives Report Positive Quantified AI Returns Yet? HKUST’s Prof. Joon Nak Choi Explains the J-Curve

Stanton Chase’s survey of 214 global executives found that 81.8% did not report a positive quantified return from AI, and HKUST’s Prof. Joon Nak Choi argues those missing returns are right on schedule. 

Foreword by Victor Filamor, Partner, Stanton Chase Greater China

When our global survey of 214 C-suite executives and board members revealed that 81.8% could not yet point to a positive quantified return from their AI investments, the finding raised an obvious question. Is the technology failing, or is something else going on? Rather than answer it ourselves, we invited someone better placed to do so. 

Prof. Joon Nak Choi of the Hong Kong University of Science and Technology works at the meeting point of AI, business analytics, and leadership, with previous appointments at Stanford University and NYU Shanghai. Few people are better positioned to interpret what our respondents told us, and his answer matters for every board currently weighing whether to press on with AI transformation or pull back. 

His argument, in short, is that the missing returns are right on schedule. AI investments follow a J-curve, and the organizations that understand this will track their progress differently, report it differently, and hold their course while others lose patience. On behalf of Stanton Chase Greater China, I am delighted to share his commentary with you. 

HKUST Adjunct Associate Professor Explains Why 82% of Executives Cannot Yet Measure AI’s Return

Commentary on Stanton Chase Survey: AI in the C-Suite: Insights from 214 Global Executives

Joon Nak Choi 
Adjunct Associate Professor 
The Hong Kong University of Science and Technology 

Stanton Chase recently surveyed 214 C-suite executives and board members on the progress of AI transformation within their organizations. Most of the results align with broader findings across the industry: the majority of represented organizations remain in an experimental phase, running pilot projects without commitment rather than embedding AI systems into their core operations. 

One finding, however, merits closer attention. When asked whether AI has delivered measurable positive returns, only 6.6% of respondents reported that they have observed a significant positive financial impact, with an additional 11.6% able to quantify modest gains. The remaining 81.8% indicated they could not yet measure any impact, that it was too early to assess, or that the impact was negative. For AI transformation teams operating under mounting pressure to demonstrate positive ROI, this gap between investment and demonstrable return has become a significant source of organizational stress. 

Three Possible Explanations

Why have so few organizations been able to demonstrate ROI? Three explanations warrant consideration. 

The first is that AI simply cannot deliver a positive return. I find this unlikely, partly because I have firsthand experience observing AI delivering substantial value, when implemented thoughtfully with appropriate support. I also find this explanation unlikely because only 1.1% of respondents reported that AI has had a negative impact on ROI—if the technology were fundamentally value-destroying, we would expect this figure to be substantially higher. 

The second possibility is that AI is already delivering ROI, but that return is difficult to capture through existing measurement frameworks. This explanation also feels implausible to me, however. If meaningful returns were materializing, we would expect to see corresponding effects in revenue growth or cost reduction—outcomes that are, by definition, visible in P&L statements. 

The third possibility is that AI requires time to deliver ROI. This explanation is, in my view, the most plausible. The concept is a familiar one in managerial economics. Transformation initiatives require initial investment and will take time to mature before they can deliver positive results. Thus, the transformation initiative will drive financial performance downward before they start delivering on revenue growth or cost reductions. Because this pattern traces the shape of the letter J, we call it the J-curve. 

Let’s trace three different curves corresponding to three different AI transformation scenarios: 

The blue dotted line in this diagram represents what senior management might expect from an AI-driven transformation initiative. Investments in AI immediately deliver productivity increases, which should be reflected by positive financial ROI. This scenario is unrealistic, unfortunately, since productivity and financial impact will almost certainly lag investment. There will be a decline in productivity before a rise. 

The red J-curve in this diagram is more realistic, as it reflects a fundamental truth on technology adoption. Achieving ROI from AI requires investment not only in the technology itself, but also in the people and the processes necessary to leverage that technology effectively. These complementary investments—in talent, training, workflow redesign, data governance, and change management—are prerequisites for returns, and they take time to mature. 

Effective change management can reduce the time required for AI investments to reach positive ROI. The dotted green line represents the realistic best case for AI implementation. Tracking this curve instead of the red line can mean the difference between continued executive support and a derailed transformation initiative. 

Implications for the C-Suite

The J-curve carries an important implication for how boards and executive teams evaluate AI transformation. Because financial metrics are a lagging indicator, they are a poor instrument for assessing how close an organization is to realizing ROI. A leadership team that waits for ROI to appear in the income statement is, in effect, looking backward rather than forward. 

Rather than relying solely on financial outcomes, AI transformation teams should track progress across three reinforcing dimensions: 
 

• Technology: the maturity and adoption of AI tools and infrastructure 
• People: the depth of talent, training, and capability being built around those tools 
• Processes: the extent to which workflows, governance, and operating models have been redesigned to capture AI’s value 

Progress across these dimensions serves as a leading indicator of the financial returns that will follow. By reporting against them, transformation teams can demonstrate that they are making progress towards positive ROI–buying them the time needed to achieve this outcome. 

______________________________________________________________________ 

About Joon Nak Choi

Joon Nak Choi is an Adjunct Associate Professor in the Department of Management at the Hong Kong University of Science and Technology, where his work spans AI, business analytics, leadership, and education. He has held visiting appointments at Stanford University and NYU Shanghai, and has played a central role in AI education and AI literacy initiatives at HKUST. He is the founder of Learnovate, an HKUST spin-off applying AI to education, and a co-creator of The AI Workforce Engine™, endorsed by the AI for Human Flourishing think tank hosted by the Harvard Flourishing Program. 

About Victor Filamor

Victor Filamor is a Partner at Stanton Chase Greater China, serving as the Regional Sector Leader for Consumer Products and Services in Asia Pacific. With 25 years of corporate experience across the Asia Pacific region and over 15 years as a retained executive search consultant, Victor has successfully placed numerous senior management and C-suite executives throughout Asia. A certified professional coach specializing in leadership and career transitions, Victor holds an MBA in Marketing from Greenwich University in Hawaii and graduated cum laude with a B.Sc. in Chemistry from the University of the Philippines, where he topped the National Chemistry Licensure Board Examinations.