Embracing the Future: AIX Society Applauds Hong Kong’s 2026 Budget and Pledges to Co-create a New AI Era
AIX Society applauds the Hong Kong AI Budget 2026, welcoming HK$100M+ in AI investments, an
When ChatGPT serves over 800 million users every week, this number already goes far beyond “technology adoption” alone. It points to an irreversible flywheel effect: once consumers normalize AI as part of daily life, the question for enterprises is no longer whether employees are willing to use it, but whether the organization itself is ready to be reshaped.
The State of Enterprise AI 2025 shows a pattern consistent with past general‑purpose technologies, from the steam engine to semiconductors. The real economic value did not appear when the technology was invented, but when companies managed to turn a raw capability into scalable, repeatable, and institutionalized use cases. Enterprise AI is now entering exactly this historical stage.
Yet this is also where most companies are most likely to take the wrong path.
The report uses large‑scale real usage data to draw a grounded picture of enterprise AI. Its analysis rests on two key sources: de‑identified usage data from over one million enterprise customers, and a structured survey of nearly 100 companies and 9,000 employees.
The results show that the shift in enterprise AI adoption is not simply “more people using it”, but that usage itself is changing in nature.
Over the past year, weekly message volume in ChatGPT Enterprise has grown by about 8×, while average per‑employee usage is up roughly 30%. More tellingly, structured workflows (such as Projects and custom GPTs) have grown nearly 19× in a single year. This indicates that employees are moving from ad‑hoc queries to reusable, shareable, and standardized AI work patterns.
At the same time, average reasoning token consumption per company has increased by around 320× over 12 months, suggesting that higher‑order reasoning models are no longer confined to experiments, but are being systematically embedded into products, services, and internal processes. AI’s role is shifting from “assistive tool” to “infrastructure layer”.
One shift that leadership should pay special attention to is how OpenAI Enterprise is positioned inside leading organizations.
In these organizations, AI is no longer “a personal assistant each employee uses on their own”, but is gradually becoming a shared language for cross‑functional collaboration and internal processes. Through project‑style workflows, custom GPTs, and APIs, companies are starting to embed policies, knowledge, and procedures into AI so they can be executed repeatedly and consistently across teams.
In practice, this means AI is becoming a new collaboration layer: a digital intermediary between people and systems that converts what used to rely on experience, word‑of‑mouth, and interpersonal tacit knowledge into machine‑readable, verifiable, and reusable flows.
This also explains why structured workflows are growing far faster than single queries. The true value lies not in asking more questions, but in turning a single successful approach into a capability the entire organization can reuse again and again.
This shift towards an AI collaboration layer is still far from widely understood in Hong Kong companies.
In recent months, when introducing AI‑centric internal collaboration and “digital employee” models to different organizations in Hong Kong, a very consistent meeting pattern keeps appearing: the people in the room are typically programmers, system architects, and IT leaders. Their questions are highly professional—how to connect APIs, how to set permissions, where data is stored, whether it complies with security and regulatory requirements.
None of these questions are wrong, but they reveal a deeper reality: in most Hong Kong enterprises, AI is still instinctively categorized as “an IT matter”.
The ones missing are usually heads of operations, HR leaders, and business line owners—the people who decide how processes run, how performance is measured, and how departments collaborate every day. The result is that AI is locked into technical discussions long before it has a chance to enter the operational core.
The enterprise AI report makes it clear that the companies truly ahead are not those with the strongest IT teams, but those that treat AI as part of their operating and collaboration capability, not just another system tool.
The nature of IT is to ensure stability, compliance, and controlled risk. Yet AI’s real value lies in process redesign, role redefinition, and the redistribution of decision rights. This is a management and organizational design transformation, not another system‑integration project.
If AI is discussed only at the IT architecture layer, it will at best improve local efficiency. Only when AI reaches the level of workflows, collaboration models, and performance logic does it begin to rewrite the company itself.
Another key finding from the report is that AI is not just substituting labor; it is expanding role boundaries.
Seventy‑five percent of surveyed employees say AI has improved the speed or quality of their work, saving on average 40–60 minutes per day, with heavy users saving more than 10 hours per week. Equally important, 75% also say AI enables them to perform tasks they previously could not do at all.
In non‑technical functions, AI‑related coding usage has increased by about 36%. This suggests AI is shortening the distance between “intent” and “execution”—professional barriers are no longer the biggest limiter. The real question is whether the organization allows capabilities to spread horizontally.
OpenAI releases new capabilities almost every three days, yet the report makes it explicit: what constrains enterprises today is no longer model performance, but whether the organization is prepared to absorb these capabilities.
For Hong Kong companies, the real risk is not “moving too slowly”, but moving in the wrong place. When AI is confined to the IT department, the company is merely upgrading tools. When AI becomes part of the collaboration and operating layers, the company is upgrading itself.
AI is not the next IT project; it is an operating‑system swap for the enterprise. This is not a short sprint, but an ongoing elimination round already in progress.
AIX Society applauds the Hong Kong AI Budget 2026, welcoming HK$100M+ in AI investments, an
Customer service is often the difference between a loyal customer and a lost one. Yet
Creating, managing, and sharing documents has always been central to enterprise operations. But in a
HKOEA Consultancy
Licence No. 59684
© 2025 Artificial Intellegence Application Research Society. All rights reserved
by Originals Group – Hong Kong, Malaysia, GBA