Corporate Transformation in the Age of AI

Over the past two years, the development of artificial intelligence (AI) and robotics has advanced at an astonishing pace. The breadth and speed of this disruption are greater, and harder to ignore, than any previous technological transformation. Studies suggest that while many organizations have already adopted AI in their daily operations, only a minority have successfully translated it into tangible business outcomes. The difference lies not in whether AI is being used, but in whether companies are willing to redesign processes, invest resources, and drive change from the leadership level.

For the construction industry and indeed all sectors, AI is not merely a technological upgrade for individuals. It represents a strategic transformation that will redefine corporate competitiveness.

Knowledge Transformation: When AI Understands Field Experience

AI has already mastered general knowledge. Domain knowledge, however, consists of specialized expertise and tacit wisdom accumulated over many years and has long been a key competitive advantage across industries. As AI agents become capable of autonomous memory, tool selection, step-by-step execution, and collaboration with one another, the way AI absorbs and utilizes organizational knowledge is fundamentally changing.

As AI's understanding and simulation of the physical world continue to improve through increasingly sophisticated world models, and as robots, sensors, and body cameras collect ever more on-site data, experiences that were once difficult for seasoned professionals to articulate will gradually become knowledge that AI can learn, replicate, and pass on. Industries such as media, information technology, and legal services have already felt the impact. Although construction, with its reliance on site-based judgment and hands-on craftsmanship, may be less susceptible to complete replacement in the near term, it will not be immune to these changes.

Organizational Transformation: From Individual Productivity to Enterprise Intelligence

If AI merely enables individual employees to work faster, its impact remains limited. An employee may complete a report 30 minutes sooner, but if approval processes and cross-departmental coordination remain unchanged, the bottleneck is simply pushed further downstream.

The real opportunity lies in enhancing Enterprise Intelligence across the entire organization. By leveraging the company's accumulated domain knowledge to develop tailored AI solutions and redesign workflows, employees can use AI to support tasks that previously depended heavily on the experience of senior personnel. At the same time, experienced professionals can focus on higher-value decision-making. This reduces organizational dependence on a small number of key individuals and helps remove genuine workflow bottlenecks.

Governance Transformation: When Robots Enter the Front Line

As AI moves from computers into the physical world, the next challenge organizations must face is robotics.

The Chinese Mainland has already established humanoid robot training centers where robots learn human movements through demonstrations and data. Yet for robots to enter frontline construction environments, the greatest challenge may not be technology itself, but liability and accountability. If a robot collides with a worker and causes an injury, or accidentally harms someone while transporting materials, who should bear responsibility?

A recent Legislative Council inquiry highlighted that Hong Kong currently has no dedicated legislation governing robot applications, nor any mandatory third-party liability insurance requirements. In recent months, incidents have included an IT professional being injured by a robot and a cyclist in Tseung Kwan O narrowly avoiding an accident involving a robotic dog. Recognizing these emerging challenges, the Department of Justice established an interdepartmental working group in March this year to review the legal framework required for AI and robotics applications.

Construction sites are far more complex than public spaces. Governments, industry stakeholders, and insurers may need to jointly develop guidelines for robotic construction operations, establish clear liability frameworks, and introduce corresponding insurance arrangements. Such measures would provide pilot users with legal certainty and appropriate risk protection. Otherwise, early adopters will have to shoulder legal risks alone, potentially slowing adoption across the entire industry.

Talent Transformation: From Operational Skills to AI Collaboration Capabilities

The first jobs affected by AI are often knowledge-based roles. Administrative work, design, and data analysis have already begun to experience significant disruption. Operational and site-based skills may be more difficult to automate, but over the long term they too will be affected. The difference is one of timing, not immunity. No occupation can be considered permanently safe.

Over the next several years or even the next decade, discussions on workforce transition, training content, and future skill requirements should begin as early as possible.

The Construction Industry Council has proposed three training priorities: enhancing frontline skills, providing professional training, and developing AI simulation and sandbox environments. Take painting robots as an example. Workers may transition from physically performing painting tasks to operating and supervising robotic systems. The jobs themselves are not necessarily being eliminated; rather, it is outdated skill sets that risk becoming obsolete.

Following this logic, the industry may need to cultivate a new generation of engineering professionals equipped with AI collaboration capabilities. Their core competencies will include asking AI the right questions, validating outputs, and ultimately taking responsibility for critical decisions. Surveys also highlight ongoing concerns regarding AI accuracy, risk management, and accountability, while identifying "Human-in-the-Loop" practices as a key characteristic of high-performing organizations.

This aligns closely with the spirit of the engineering profession's certification and sign-off system. Critical decisions should not be delegated entirely to tools. Instead, qualified professionals must remain responsible for oversight, while every action taken by AI systems and robots should be supported by traceable records and emergency shutdown mechanisms.

AI-driven transformation will not pause simply because organizations hesitate. Rather than waiting until every question has been answered, companies should start today by creating environments where employees can experiment with AI firsthand. However, experimentation must go beyond individual usage. Organizations need to systematically document successful use cases, lessons learned from failures, risk-control measures, and optimized workflows. Only then can AI evolve from a collection of standalone tools into a genuine organizational capability.