BUBBLE?
Only a handful of companies are seeing rapid revenue growth and higher valuations, and it is not enough to call the AI trend a transformation, an analyst said
-
By Crystal Hsu /
Staff reporter
Artificial intelligence (AI) is entering a more challenging phase next year as companies move beyond experimentation and begin demanding clear financial returns from a technology that has delivered big gains to only a small group of early adopters, PricewaterhouseCoopers (PwC) Taiwan said yesterday.
Most organizations have been able to justify AI investments through cost recovery or modest efficiency gains, but few have achieved meaningful revenue growth or long-term competitive advantage, the consultancy said in its 2026 AI Business Predictions report.
This growing performance gap is forcing executives to reconsider how AI is deployed across their organizations, it said.
Photo: Reuters
“Many companies are getting something back from AI, but not enough to call it transformation,” PwC Taiwan chief digital officer Chang Chin-jui (張晉瑞) said.
Only a handful of companies are seeing rapid revenue growth and higher valuations, while the majority experience limited returns, and measurable productivity gains remain difficult to pin down, he said.
In many cases, AI investments break even or deliver incremental benefits rather than drive true transformation, he said.
The dynamics are starting to change as corporate leadership plays a more active role, the consultancy said.
PwC said it expects firms to move away from scattered pilot projects toward top-down AI strategies in the coming year.
Senior executives are increasingly focusing on a small number of high-impact processes, directing resources to initiatives that promise the greatest value rather than funding dozens of disconnected projects, it said.
Centralized AI studios are also becoming more common, giving companies tighter control over technology, talent and operational standards, it added.
A key area of focus would be “agentic AI” — systems capable of carrying out tasks autonomously, PwC Taiwan said.
With clearer performance targets, stronger testing and tighter oversight, these AI systems are expected to become more effective and reliable next year, it said.
The shift is also poised to reshape the workforce. As AI takes over routine and technical tasks in areas such as software development, finance and administrative operations, the demand for mid-level specialists is likely to decline, it said.
Companies are expected to prioritize employees who understand business processes and can supervise AI systems, allowing senior staff to focus on strategy and decisionmaking, it said.
This change is anticipated to increase demand for multidisciplinary “generalists,” who can oversee agentic AI and ensure its actions remain aligned with corporate goals, it said.
Another emerging priority is AI orchestration — the process of integrating multiple AI models, tools, data sources and workflows to function seamlessly as a unified system, PwC Taiwan said.
These control layers would be crucial for turning conceptual ideas into operational production systems, while maintaining security, compliance and oversight, it added.
link