HK’s AI Moment: 91% of Companies Doubling Down, But ROI Question Remains Unanswered

Forget the AI bubble talk. Hong Kong’s C-suite has already moved on.

According to Accenture's recently released Pulse of Change survey, which involved 3,650 C-suite leaders, 3,350 employees, 20 industries, and 20 countries, a staggering 91% of Hong Kong organizations plan to increase their AI investments in 2026. If the so-called AI bubble actually bursts? More than half (51%) would still ramp up spending, and 60% would keep hiring. That's conviction.

But conviction without clarity is just expensive noise. And therein lies the Hong Kong CDO’s existential problem for 2026.

The agents are already in the building

Hong Kong isn't just talking about agentic AI; it's already deploying it. 37% of organizations are actively rolling out AI agents across multiple business functions, outpacing both global averages and their mainland China peers. Process redesign is happening at 33% of companies, and 28% of C-suite leaders claim to use generative AI tools daily.

Robert Hah, strategy and consulting lead and managing director for Accenture Greater China's Hong Kong office, framed the stakes: “In the beginning of the new year, we are seeing a series of positive and unique trends among Hong Kong companies and their employees. While adopting a more pragmatic approach, they continue to explore innovation – this will enable them to better navigate multiple challenges, boost resilience and competitiveness, and further deploy technologies like artificial intelligence as a catalyst for business growth. It is increasingly clear that only by bringing their people along the journey can companies realize true AI's values and translate the optimism-led momentum with generating long-term value across the organization.”

That last line is the one CDOs should tattoo on their roadmaps.

The hidden truth: Nobody's measuring the right things

Here's where the rosy and optimistic picture gets uncomfortable. Only 11% of Hong Kong leaders cite ROI as the primary driver of AI investment. The top motivators? Risk management confidence, and competitive positioning. Worthy goals, but they're not a business case. They're a defensive crouch dressed up as strategy.

Eight in 10 Hong Kong C-suite leaders say their experience with generative AI has unlocked greater business impact potential. Yet only one in four strongly agrees that AI is helping their organizations shift from measuring activities to measuring true outcomes. That gap between perceived potential and demonstrated value is exactly where CDO careers go to die.

The data architecture question isn't glamorous, but it's foundational. C-suite leaders themselves acknowledge that having the right data strategy and core digital capabilities in place is the single most important accelerant for AI scaling. If your data estate is a swamp, your AI agents are just expensive swamp monsters.

The yawning employee trust deficit

The survey delivers a split-screen moment on workforce sentiment. Over half of Hong Kong employees report that AI has genuinely improved the quality of their work. Yet 45% also say AI tools have produced low-quality or misleading outputs, leading to wasted time and effort. Both things are true simultaneously, and that cognitive dissonance is exactly what makes AI adoption so hard to govern.

Add to that: only 19% of Hong Kong employees strongly agree they'd comfortably delegate tasks to an AI agent, compared to 28% on the mainland. And 55% reach for an AI tool before asking a colleague, but that number still trails China's average of 64%.

Less than one in five employees strongly agree that their organization rewards curiosity and experimentation with AI. When you're asking people to co-create with machines, that's a cultural failure with a direct data consequence.

The CDO’s next steps

The survey paints a picture of an ecosystem that's fast but not yet deep. For CDOs navigating this landscape, three pressure points demand immediate attention:

The measurement architecture needs to be rebuilt around outcomes, not outputs. Tracking AI tool adoption is vanity. Track decision velocity, error reduction, and revenue attribution instead.

The data sovereignty agenda is moving up the priority stack. Hong Kong companies are increasingly focused on operational control, supply chain resilience, and data ownership. All of these require CDOs to own the governance conversation before regulators do.

And the trust gap between leadership enthusiasm and employee reality must close. That means feedback loops, visible iteration, and yes, actually listening when the tools fail.

The AI wave in Hong Kong is real. But waves don't build companies; infrastructure does.

Image credit: iStockphoto/treety