PTS gives practical, vendor-neutral AI advice to businesses in Hong Kong and Mainland China — starting with the job you’re trying to do, not the tool, then matching it to what’s actually accessible and safe for your data where you operate.
Most AI advice starts with the wrong question: “should we use ChatGPT, or Claude?” In Hong Kong and China that question can’t be answered in the abstract, because two things decide everything — where you operate (the tools available differ sharply, and some aren’t available at all) and how sensitive your data is. We make those the gate. The result is a short, honest path from a real business problem to a tool you can actually use, safely.
We are technologists, not your lawyers: we operationalise AI — tool selection, data handling, governance and rollout — alongside your legal advisers, and to our own ISO/IEC 27001 and ISO/IEC 20000 disciplines. Not sure where to start? Try the 30-second check above, then read on. Useful background: AI tools compared — what works in Hong Kong & China, can you use ChatGPT in Hong Kong? and the Hong Kong & China data laws guide.
Start with the job, not the tool
The businesses that get value from AI start from a task that is slow, repetitive or knowledge-heavy — not from a product. Pick two or three concrete jobs with measurable value, and the right tool follows. Common ones for SMEs:
Writing & communications
Drafting emails, proposals, reports and first drafts — the highest-volume, lowest-risk place most teams start.
Summarising & research
Turning long documents, threads and reports into a usable brief in seconds — a genuine time-saver for busy teams.
Knowledge & document search
Answering “where’s the policy on…?” from your own documents — best done grounded on your data so answers stay inside your boundary.
Customer service
Drafting replies, deflecting routine questions, or a front-line chatbot — high value, but a public-facing bot raises extra questions in the Mainland.
Translation
EN ⇄ 繁 ⇄ 简 at speed across a cross-border team — a natural fit for Hong Kong and China operations.
Meeting notes & analysis
Transcribing and summarising meetings, and first-pass analysis of spreadsheets — useful, with care over who and what is recorded.
The two questions that decide everything: data and place
Before any tool is chosen, two questions set the boundaries. They are the difference between AI that helps and AI that creates a problem.
Classify the data
How sensitive is the data the task touches — public, internal, confidential, or personal/regulated? The more sensitive, the narrower your safe options. The single rule that prevents most mistakes: never put confidential, personal or regulated data into a free consumer chatbot, whose terms may allow your inputs to train the model. For anything sensitive, use a tier that contractually won’t.
Know where you operate
Where do the users and the data sit — Hong Kong, Mainland China, or across the border? This decides what’s even available (some tools are blocked or unsupported) and which rules apply to moving data. The two tracks below set out each reality.
Match, then govern
Only once data and place are clear does the tool choice make sense — and the tool is only half of it. Sensitive use needs governance around it: a written policy on what staff may and may not put in, the right accounts, and a record of what’s happening. We design that with you.
Using AI in Hong Kong
Hong Kong has no AI-specific law; the technology-neutral Personal Data (Privacy) Ordinance (PDPO) governs what data you may put into AI tools, and the Privacy Commissioner (PCPD) has published genuinely useful, practical guidance you can operationalise.
What’s accessible. ChatGPT and Google Gemini are reachable directly from Hong Kong. Anthropic’s Claude is not available first-party — Hong Kong isn’t a supported market, so direct business access runs only through intermediaries (Copilot, Perplexity and the like). That’s a commercial/region restriction, not a Hong Kong law.
The practical safe route for real business work is Microsoft 365 Copilot and Copilot Chat, signed in with a work account. These carry enterprise data protection: your prompts, responses and the data they reach through Microsoft Graph are not used to train the underlying models, Microsoft acts as your data processor under contract, and the protection is included in Copilot Chat at no extra cost. That contractual “won’t train on your data” commitment is exactly what a consumer chatbot doesn’t give you — and it’s what makes Copilot the sensible default for confidential work. (Two things to know: the protection attaches to the work sign-in, not personal accounts; and web-search-grounded queries are handled under separate terms.) For organisations standardised on Microsoft, Microsoft Purview can even bring some third-party AI tools under governance from inside your tenant.
Follow the official guidance. The PCPD’s AI: Model Personal Data Protection Framework (2024) and its Checklist on Guidelines for the Use of Generative AI by Employees (2025) tell organisations to give staff clear instructions on what may and may not be entered into AI tools — naming personal, confidential, proprietary and copyrighted data as categories to restrict — and to anonymise or cleanse personal data before input. That’s best practice we turn into a working policy and the right tooling. See our Microsoft 365 & cloud and Microsoft 365 security services for the platform side.
Using AI in Mainland China
Mainland China is a different world, and pretending otherwise is how projects go wrong. The major Western tools — Claude, and others — are blocked or unreliable behind the Great Firewall, and China runs its own AI and data regime. There are three realistic routes, chosen by use case and data:
- A compliant domestic model. China has a deep field of capable local models with enterprise and API offerings. For China-facing work on China data, an in-country model is usually the right answer.
- A governed cloud inside China. Azure operated by 21Vianet is a physically separated, China-sovereign instance of Azure — a route to keep AI workloads and data inside the Mainland under local operation. (Which specific AI services it carries changes; we confirm current availability per engagement.)
- Keep the data and workload outside China for cross-border-permissible use cases — often the cleanest option when the data doesn’t legally need to stay in the Mainland.
Two points matter most for SMEs. First, purely internal AI use generally sits outside China’s heaviest regime: the public-service security-assessment-and-filing obligations attach to generative-AI services offered to the public in China — using AI inside your own business is treated differently. Second, moving data out of the Mainland is a regulated cross-border transfer, with thresholds that decide whether you’re exempt, need a filing, or need a full security assessment. That’s the part to get right before you pipe China data into a tool hosted elsewhere — our Hong Kong & China data laws guide (with a self-check) walks through it, and our China IT services team operationalises it on the ground.
How we run an AI engagement
A short, deliberate path from problem to a working, governed result — and you can stop after any stage.
1. Use case. We start from the jobs to be done and the value each creates, and pick the two or three worth doing first.
2. Data & place. We classify the data sensitivity and map where users and data sit — the gate that rules tools in or out.
3. Tool fit. We recommend the approach that’s accessible, appropriate and safe for that combination — vendor-neutral, on merit, not whatever we’d profit from.
4. Pilot. We run a small, measured pilot with guardrails — an acceptable-use policy, the right accounts, and a human in the loop — and prove the value before you commit.
5. Scale & govern. We roll out, train your people, and put governance around it — then revisit, because access and rules move fast.
Want to use AI without creating a data problem?
Practical, vendor-neutral AI advice · Hong Kong · Mainland China
Why PTS for AI advice
Plenty of people will tell you which AI tool to buy. Fewer can tell you what works where you operate, keep your data out of trouble, and then make it run.
Vendor-neutral
We don’t sell a model or a platform, so the advice serves your use case — not a licence we’re reselling. The right answer is sometimes “you don’t need a new tool at all.”
Cross-border by default
We work in both Hong Kong and Mainland China every day. We know what’s accessible on each side of the border and how data is allowed to move between them — the gap most generic AI advice ignores.
Data-first and governed
We lead with data sensitivity, not hype — and we run to our own ISO/IEC 27001 and ISO/IEC 20000 disciplines, so the governance, access control and policy around your AI use are real, not an afterthought.
Practical, and we run it
We’re not a slide deck. We pilot, deploy and then support the result day to day as part of managed IT services — operationalising the law alongside your advisers, not just talking about it.
AI advisory FAQs
Straight answers to what Hong Kong and China businesses ask us most about adopting AI.
Can I use ChatGPT and Claude in Hong Kong?
ChatGPT and Google Gemini are reachable directly from Hong Kong. Anthropic’s Claude is not available first-party — Hong Kong isn’t a supported market, so business access is only through intermediaries such as Microsoft Copilot or Perplexity. For confidential work the better route in any case is an enterprise, data-protected tier (e.g. Microsoft 365 Copilot signed in with a work account), not a consumer app.
Can I use AI tools in Mainland China?
The major Western tools (Claude and others) are blocked or unreliable in the Mainland. Realistic options are a compliant domestic model, a governed cloud inside China (Azure operated by 21Vianet), or keeping the workload and data outside China where the use case allows. Which is right depends on the data and the job — that’s the advice we give.
Is it safe to put company data into AI tools?
It depends entirely on the tool tier and the data. The rule we give every client: never put confidential, personal or regulated data into a free consumer chatbot, because its terms may allow your inputs to train the model. For sensitive work, use a tier with a contractual no-training, enterprise-data-protection commitment, and put governance around it.
Do you sell a particular AI platform?
No — we’re deliberately vendor-neutral. We recommend the approach that fits your use case, jurisdiction and data on merit. Often that’s a tool you already pay for (such as Microsoft 365 Copilot); sometimes it’s no new tool at all.
Do you give legal or compliance advice?
No. We’re your practical technology advisor — we handle tool selection, data handling, governance and rollout, and we operationalise the rules alongside your own legal advisers. For the legal specifics of the privacy and cross-border rules, take qualified Hong Kong or PRC advice; our data laws guide is a practical starting point, not legal advice.
Can you help with AI across both Hong Kong and China?
Yes — that’s our distinctive ground. We advise on what’s accessible and appropriate on each side of the border, and how data may move between them, with engineers on the ground in both. See China IT services.
What AI use cases actually pay off for an SME?
Usually the unglamorous ones: drafting and communications, summarising long documents, searching your own knowledge, translation across a cross-border team, and meeting notes. We help you pick the two or three with the clearest payback and prove them in a pilot first.
How much does AI advisory cost?
It depends on scope — a focused pilot is a different engagement from an estate-wide rollout with governance. We don’t publish fixed fees; we scope your needs and put a clear, costed proposal in writing. Talk to PTS to start.
Related services
AI sits on top of the rest of your technology. PTS also delivers:
- Microsoft 365 & cloud — the Copilot and Microsoft 365 environment most safe AI use runs on.
- Microsoft 365 security — data governance and protection (including Purview) for AI-ready tenants.
- China IT services — AI and IT on the ground in the Mainland, where the rules differ.
- Hong Kong & China data laws — the cross-border and privacy reality, with a self-check.
- Managed IT services — running and supporting it all day to day.