5 strategic steps for HK businesses to unlock AI value

Only 3% of Hong Kong businesses have fully integrated AI into their workflows, citing hurdles including resistance to organisational change and outdated digital infrastructure, according to Accenture’s inaugural Hong Kong AI Index report.

Surveying 150 companies across nine industries between March and April 2026, the study revealed a significant gap between corporate ambitions and operational reality. Over the next 18 months, business leaders in the city expect scaled AI deployments to deliver a 15% increase in productivity, a 14% uplift in revenue, and an 11% reduction in operational costs. However, only 12% of organisations’ strategic AI initiatives have reached scale, while just 3% have fully integrated AI and data into their core business strategies.

A elite group of organisations identified in the report as "Vanguards" are outperforming their peers by scaling 22% of their generative AI investments. These frontrunners are twice as likely to report returns that exceed expectations and demonstrate a stronger ability to identify high-value opportunities.

Samantha Zhu, chairperson of Accenture Greater China, said, "Chinese enterprises have accelerated paces embracing AI holistically as the country is looking to AI as a key force to propel high-quality development. Yet to turn early impact into lasting value, they need a systematic approach that focuses on strategy, technology, process, organisation and governance to enhance efficiency and drive growth."

Robert Hah, lead of Hong Kong, Accenture Greater China, said, "Hong Kong has all the ingredients it needs for AI-led growth. The city has strong institutions, deep capital markets and privileged access to both China's AI ecosystem and global innovation networks, forming a distinctive position to act as both a 'super connector' and 'super value-adder'."

"It's imperative for companies to move beyond experiment to embedding AI across the whole operations, better addressing challenges such as shrinking workforce and slowing productivity growth, and further capitalising on the city's AI + Initiative - industries for AI and AI for industries," he added.

Addressing concerns over AI-driven job displacement, Hah told MARKETING-INTERACTIVE that increased AI adoption actually creates new employment opportunities rather than simply eliminating roles. As an example, he noted that Accenture itself is actively recruiting roles such as talent futurists to analyse evolving skill requirements and help upskill existing employees.

The report identifies structural and organisational barriers as primary constraints to progress. Over the coming 18 months, local firms anticipate significant maturation, with expectations that upper-level large language model operational capabilities will rise from 3% today to 47%, and robust data management capabilities will jump from 24% to 91%. Nevertheless, 45% of respondents warned that stagnant operating models remain a major bottleneck to productivity.

To help organisations bridge this execution gap, Accenture laid out four strategic imperatives for sustainable AI integration:

1. Lead with value so AI investment follows business ambition

AI creates enterprise value when it delivers measurable outcomes across growth, productivity, customer and employee experience, talent and sustainability, and when those outcomes scale beyond isolated use cases and endure over time. Most Hong Kong firms are not yet there. Only 3% have fully integrated AI and data into core business strategy. Just 34% report consistent AI value across multiple functions or widespread sustained value across the enterprise. The bottom line? Most companies are failing to connect AI use cases to what the business is actually trying to achieve.

Accenture suggested that companies must reframe business priorities through an AI lens, concentrate investment around strategic outcomes and redesign how decisions are made, experiments are funded and pilots earn the right to scale. Hah added that this step is foundational, as organisations must clearly define their goals before equipping employees.

2. Reinvent work so people and AI create value together

Work environments are changing more slowly than the underlying technology. Only 19% of Hong Kong firms view process reinvention as critical to scaling AI, and just 26% plan to redesign roles. Furthermore, while 78% of leaders express confidence in attracting talent, only 15% possess the specialist AI skills needed today.

That is why redesigning work comes before aligning skills. Executives should allocate tasks across human judgment, copilots, agents and automation, then reassemble them into roles and teams with clear accountability and oversight.

3. Bridge the AI skills gap and modernise talent strategy

Closing the persistent skill gap demands a modern talent and upskilling strategy that evolves alongside technological change. To address the shortage of specialist expertise, enterprises should leverage workforce intelligence to continuously monitor emerging skill requirements and identify obsolete practices.

Integrating role-specific, experiential learning directly into daily employee routines creates an adaptable culture where human talent and artificial intelligence capabilities grow together, mitigating displacement concerns through proactive internal career pathing.

4. Rewire the digital core to scale agentic AI

Enterprise-wide AI integration remains in its early stages, with just 1% of surveyed firms deploying agentic AI across functions. Inflexible legacy infrastructure was cited by 41% as a primary hurdle, compounded by increasing system complexity as 57% of firms run models from multiple provider groups across Mainland China, Hong Kong, and international markets.

A digital core capable of scaling agentic AI needs to do three things. Connect AI capabilities across functions. Build a semantic layer—underpinned by domain ontologies—that gives people and AI a common understanding of the business and makes the right information available in the right context. Orchestrate different models based on performance, cost, capability and compliance, while modernising legacy systems that constrain flexibility and scale.

5. Embed RAI so trust grows alongside scale

Trust must be integrated into how AI is designed, governed, and monitored. Only 13% of local firms report high levels of RAI capability today, and RAI currently receives just 13% of average data and AI budgets.

Building RAI into the operating machinery starts with model selection. Enterprises should assess providers' evaluation practices, safeguards, transparency and incident-response arrangements before deployment, and maintain oversight as models evolve. Risk assessment, testing, explainability and continuous monitoring must permeate the entire AI lifecycle. As systems become more agent-driven, clear decision rights should identify who approves an application and who remains accountable for its outcomes. Technical architecture should accommodate regulatory changes so AI systems can adapt without extensive re-engineering.

For Hong Kong companies operating across the Mainland and global markets, strong RAI governance can expand the freedom to scale. It makes it easier to earn employee and customer trust, navigate different requirements and take successful applications into more parts of the business and more markets.

Elevate your brand with cutting-edge AI, emerging tech, and data-driven experiences. Join 200+ marketing leaders at Digital Marketing Asia HK 2026 on 27–28 October to uncover game-changing trends, proven playbooks, and the bold strategies shaping the future of marketing.

Related articles:

Google: 1 in 5 HK citizens are AI 'super users'
China’s generative AI user base surpasses 700m in H1 2026