70% of Hong Kong white-collar workers use AI - but top executives lag behind

70% of Hong Kong's white-collar staff use AI. Leadership still lags.

A December survey of 4,500 professionals and tertiary students found that over 70% of white-collar workers in Hong Kong already use AI on the job. Nearly 90% say it makes them more productive, and over 90% engage with AI daily.

The outlier: less than 15% of founders and executives use AI themselves. That gap slows integration and sends a mixed signal to teams.

As McKinsey partner Jackey Yu put it, "Senior management needs to be onboard…and act as a role model." Managing partner Arthur Shek added that companies need clear AI goals and real investment in skills.

What this means for managers

Your people are already using AI. Without direction, they'll create their own playbooks-good or bad.
Executive hesitation blocks budgets, tooling, policy, and measurable outcomes.
Teams want clarity: which tools, for which workflows, with what standards and guardrails.

The 90-day plan to align AI with outcomes

Define 3 business wins: e.g., reduce report prep time 40%, cut customer response time 30%, ship proposals 2x faster.
Pick 3 priority workflows per function (sales emails, research summaries, meeting notes to tasks).
Standardize tool access: approve 1-2 AI assistants and data-safe extensions; document when to use them.
Write a one-page policy: data handling, confidentiality, human review, citation standards, and banned use cases.
Name champions in each team to collect prompts, patterns, and pitfalls; share wins weekly.
Upskill fast: short courses, prompt patterns, and role-based playbooks tied to OKRs.
Instrument the work: baseline current time/cost, then track deltas on cycle time, quality, and error rates.
Lead by example: executives demo AI use in meetings and reviews; model the behavior you want.

Starter use cases by function

Sales: draft outreach, qualify leads, summarize calls, create account plans.
Marketing: brief creation, content outlines, repurposing assets, SEO ideas with human edit.
Operations: SOP drafting, variance analysis summaries, vendor email routing.
Finance: commentary on variances, plain-English budget notes, policy Q&A.
HR: JD drafts, interview guides, policy summarization, learning paths.
Product: user story drafts, test-case generation, meeting notes to action items.

Students are ready. Are you?

Over 70% of tertiary students prefer AI-enabled careers, and 90% are actively building relevant skills. Yu summed it up: "Embrace AI, no-brainer… think more about how to partner with AI to do more and better."

Update job descriptions with clear AI expectations and tools used.
Add practical AI tasks to interviews (e.g., summarize a brief, improve a prompt, design a workflow).
Offer internships or sprint projects that pair students with team AI champions.

Governance without the drag

Set data rules: what can/can't be pasted, how to mask PII, and approved connectors.
Keep a human-in-the-loop for external content, financials, legal, and safety-critical decisions.
Log prompts for QA and learning; periodically review outputs for bias and accuracy.

How to measure impact (simple and defensible)

Cycle time per task (before/after), error rate, and satisfaction score from end users.
Adoption rate by team, prompts/playbooks reused, and cost per deliverable.
Quarterly review: retire low-value use cases, double down on proven ones.

Bottom line

Your teams already use AI daily. Close the leadership gap, set clear targets, invest in skills, and make AI a visible part of how management works. The companies that do this will move faster with fewer meetings and cleaner handoffs-and that shows up in the P&L.