近日，教育工作人員總工會發表了一份關於在職教師工作壓力與身心健康的報告：本港教師整體工作壓力偏高，而最大的壓力來源竟然不是教學本身，而是非教學相關的「行政及文書工作」，壓力評分高達 6.31 分（滿分 7 分）。
當社會各界每天都在鼓吹AI 大時代、教人如何用 AI 生成圖片、做精美簡報、寫代碼等來提升生產力時，為甚麼每天被行政及文書纏擾的打工仔，卻享受不到科技的便利？
💡 打工仔的集體無奈：自學 AI 的代價與限制
香港近年確實有很多免費或付費的 AI 課程。但對大部分打工仔來說，這正演變成一場科技轉型焦慮：
- 成本轉嫁員工：要在本已飽和的工作時間外，自己付費或犧牲休息時間去自學 AI，對很多身心俱疲的打工仔來說，門檻高得令人卻步。
- 基本建設侷限 AI 發揮：即使花費心力學會了，回到辦公室卻發現不能學以致用。很多重覆性的行政工作——例如填寫或遞交繁瑣表格、人手 Filing、將數據重覆輸入到多個不同的舊 Excel 表——這類傳統工序，個人的 AI 助手顯得無用武之地。
這牽涉到科技轉型最核心的底層問題：不是員工不會用 AI ，而是機構傳統電子化與系統自動化脫節！
📊 權威數據：香港機構的數碼化脫節
作為 AI 及數碼轉型顧問，我們必須指出一個殘酷的現實。根據最新的 2026 年度工作趨勢指數報告，香港員工的個人 AI 應用速度，遠遠領先於企業系統的轉型進程。
參考香港生產力促進局的《香港企業數碼化指數調查》，當時整體企業的數碼化指數僅為 35.9 分（滿分 100 分），屬於基本起步水平，當中數據的運用分項指數更是最低（僅 32.0 分）。更重要的是，大型企業數碼化指數有 52.4 分，但中小企及缺乏獨立 IT 部門的教育/非牟利機構，平均僅得 33.9 分。（註：此為該系列最近期之 2023 年調查數據。鑑於本港非商用機構底層系統更迭緩慢，其反映的結構性滯後在 2026 年依然高度吻合。）
另一項 2025 年的調查亦指出，推行 AI 面臨的其中一大技術挑戰，正是難以與現有的舊系統整合，技術架構不兼容。甚至許多機構根本沒有系統，大多數工作流程都是口耳相傳，沒有一個人能清晰說出整個流程。
🧠 顧問觀點：沒有結構化數據，就沒有 AI 轉型
在科技界有一句名言：「垃圾進，垃圾出（Garbage in, Garbage out）」。
如果機構內部的行政表格仍然倚賴紙本交收、流程決策全憑個人判斷、數據分散在不同的數據孤島（Data Silos）或紙本中，直接引入 AI 只會加速製造垃圾。AI 需要的是結構化的電子數據（Structured Data）以及標準化的數碼工作流程（Workflows）。
要成功推動轉型必須由上而下推動，不能將責任轉嫁給員工獨自完成。 機構在擁抱 AI 前，必須先完成這三步的流程重組（Process Re-engineering）：
📥 電子化（Digitization）：將所有紙張、人手記錄徹底轉為電子數據。
⚙️ 系統化（Systematization）：打破各部門舊系統的隔閡，讓數據能自動互通。
🤖 自動化（Automation）：起用新系統或 RPA 串聯舊系統工序，消除人手重覆入數與 Filing 的煎熬。
做好這些基礎建設，AI 才能作為最後的催化劑發揮作用，真正為員工減壓，為機構創造價值。否則，AI 轉型只會一直只聞樓梯響，不見人下來。
Recently, the Hong Kong Federation of Education Workers Union released a report on the work stress and mental health of in-active teachers. The results are alarming: teachers in Hong Kong face exceptionally high levels of overall work stress. Strikingly, the primary source of this pressure stems not from teaching itself, but from non-instructional “administrative and clerical work,” which scored a staggering 6.31 out of 7.
While society aggressively hypes the “AI Era” every day—urging people to learn image generation, slide creation, and coding to boost productivity—why are back-office employees, who drown daily in forms and data, completely locked out of these technological benefits?
💡 The Collective Frustration of Employees: The Cost and Limits of Self-Taught AI
Numerous free and paid AI courses have flooded Hong Kong recently. Yet, for most office workers, this trend has mutated into a form of “digital transformation anxiety”:
- Costs Shifted to Employees: Forcing exhausted workers to spend their own money or sacrifice their precious rest hours outside of an already saturated schedule to learn AI creates a barrier that deters many.
- Infrastructure Bottlenecking AI Potential: Even after spending immense effort to learn these tools, workers return to the office only to face a harsh reality. Their new skills find no application. Repetitive administrative tasks—such as filling out tedious forms, manual filing, and duplicate data entry across multiple legacy Excel sheets—leave personal AI assistants completely useless.
This exposes the absolute core flaw of digital transformation: The bottleneck is not that employees don’t know how to use AI; it is that corporate legacy digitization and system automation are severely disconnected!
📊 Authoritative Data: The Structural Lag in Hong Kong Organizations
As AI and digital transformation consultants, we must highlight a brutal reality. According to the latest 2026 Work Trend Index report, the speed at which Hong Kong employees adopt personal AI tools vastly outpaces the digital transformation of corporate enterprise systems.
Consider the Enterprise Digitalisation Index published by the Hong Kong Productivity Council. The overall enterprise digitization index stood at a mere 35.9 out of 100, indicating a basic, entry-level stage. The sub-index for “Data Utilization” performed the worst, scraping just 32.0 points. Crucially, while large enterprises scored 52.4 points, SMEs and educational/non-profit organizations lacking dedicated IT departments averaged only 33.9 points. (Note: This data is from the most recent 2023 index survey in this series. Given that infrastructure upgrades in non-commercial organizations move slowly, the structural lag it reveals remains highly accurate in 2026.)
Another industry survey from 2025 indicated that the single biggest technical challenge in deploying AI is integrating it with legacy architecture. Systems are fundamentally incompatible. In fact, many organizations operate without any formal systems at all; the majority of their workflows are passed down entirely by word of mouth, with not a single person able to clearly articulate the end-to-end process.
🧠 Consultant Insights: No Structured Data, No AI Transformation
A golden rule in technology states: “Garbage in, garbage out.”
If internal administrative forms still rely on paper routing, if process decisions hinge entirely on individual judgment, and if data remains trapped in paper stacks or siloed databases, rushing to introduce AI will only accelerate the production of garbage. AI inherently demands structured data and standardized digital workflows.
Successful transformation must be driven from the top down; the responsibility cannot be dumped onto employees to figure out in isolation. Before blindly rushing to embrace AI, organizations must first execute a three-step Process Re-engineering:
📥 Digitization: Convert all paper documents and manual logs entirely into electronic data.
⚙️ Systematization: Break down the barriers between legacy departmental systems so data flows automatically.
🤖 Automation: Deploy new systems or RPA (Robotic Process Automation) to bridge legacy workflows, eliminating the agony of manual data entry and filing.
Only by establishing this foundation can AI act as the ultimate catalyst, genuinely relieving employee stress and driving corporate value. Otherwise, AI transformation will remain a hollow slogan—all talk and no action.
參考資料 Reference:
HKPC https://www.hkpc.org/zh-HK/about-us/media-centre/press-releases/2023/enterprise-digitalisation-index