Your Customer Data Has a Privacy Problem You Can’t Ignore

Every time you use a cloud-based AI tool to draft an email, summarize a report, or analyze customer feedback, you’re sending your business data—and your customers’ personal information—to a server that could be anywhere in the world. For a Malaysian SME owner, this isn’t just a technical detail. It’s a trust issue. Your customers expect you to protect their data. But the AI tools you rely on might be storing that data on servers in countries with different privacy laws than Malaysia’s Personal Data Protection Act (PDPA).

This is why a recent announcement from Ukrainian app developer MacPaw caught our attention. MacPaw is partnering with Liquid AI to run AI models directly on devices—not in the cloud. The goal is to let users run AI assistants and workflows offline, with better privacy and security. This isn’t just news for tech giants. It’s a signal of where AI is heading: away from the cloud and into your pocket, your laptop, and eventually, your business operations.

TL;DR: Major app developers are moving AI processing from distant servers to local devices. This shift means faster responses, stronger data privacy, and the ability to work offline. For Malaysian SMEs, it could mean using AI tools without worrying about customer data leaving your control.

What This Means: AI That Comes to You, Not the Other Way Around

Most AI tools you use today are “cloud-based.” You type a prompt, it travels over the internet to a data center, the AI processes it, and the result travels back. This works fine until you have a bad internet connection—which happens in many parts of Malaysia, from industrial parks in Shah Alam to new office areas in Johor Bahru.

On-device inference changes this. The AI model lives directly on your device. No data leaves your phone, laptop, or tablet. MacPaw is developing this exact system, called Elix, with Liquid AI. The companies are building a “local memory system” that lets AI assistants remember context without sending that information to external servers. Liquid AI’s CEO emphasizes that their models are designed specifically for the hardware they run on, making them more efficient and private.

Think of it this way. Right now, using cloud AI is like hiring a contractor who insists on taking your business plans home to work on them overnight. On-device AI is like having that contractor work at a desk in your office, where you can see everything happening. The work gets done, but nothing sensitive leaves your premises.

How This Applies to Malaysian SMEs

Malaysia’s digital economy is growing rapidly, with MDEC actively pushing digital transformation for SMEs. But many business owners we talk to are hesitant about AI adoption—not because they don’t see the value, but because they worry about data security. This local AI trend directly addresses that concern.

Consider a typical scenario: a Grab Food merchant partner or a small retail shop in Penang that processes hundreds of customer orders daily. If they use a cloud AI tool to predict inventory needs or analyze sales patterns, they’re sending transactional data to external servers. With on-device AI, that analysis happens right on their existing point-of-sale system or laptop. No third party sees your sales figures, your supplier pricing, or your customer purchase history. For businesses preparing for PWDC compliance or those handling sensitive customer information, this is a meaningful advantage.

Another example: professional services firms like accounting firms in KL or legal practices in Georgetown frequently handle client documents that are protected by confidentiality agreements. Many are prohibited from uploading client files to external cloud services. On-device AI lets these firms summarize contracts, draft communications, or search through documents using AI, all while keeping client data on their own secure servers or even air-gapped machines. This expands what SMEs can do with AI without violating professional obligations or client trust.

There’s also a practical, everyday benefit: reliability. Malaysia’s internet infrastructure, while improving, isn’t always consistent—especially during monsoon season when connectivity can be disrupted. If your AI tools work locally, a brief internet outage doesn’t stop your business intelligence function. You can still generate invoices, get customer insights, or draft proposals even when your connection is down. For SMEs operating in industrial areas where fiber infrastructure isn’t guaranteed, this resilience is valuable.

“The real value of on-device AI for Malaysian SMEs isn’t just about privacy—it’s about taking back control of your business data. When your intelligence tools work locally, you own the entire process.”

MacPaw’s approach also points to a hybrid future. The company plans to offer both local models and access to cloud models (from companies like Google) through one platform. Malaysian SMEs can expect similar options: use local AI for sensitive or routine tasks, and switch to cloud AI for heavier processing needs when appropriate. This flexibility means you don’t have to choose between capability and confidentiality.

Practical Takeaways: What You Can Do Today

You don’t need to wait for this technology to mature. Here’s how to start preparing your business for the local AI shift:

• Audit your data flows. Identify which of your current tools send business data to external servers. You might be surprised by how many apps are uploading your files.
• Ask your software vendors about local options. When renewing contracts, ask whether they offer on-device processing or private deployment options.
• Prioritize privacy-sensitive workflows. Start with AI use cases involving customer personal data, financial information, or proprietary processes. These should be first in line for local processing.
• Test offline capabilities. Before committing to a new AI tool, try it without an internet connection. This tells you whether it can handle local processing.
• Document your data handling practices. Whether you’re dealing with PDPA compliance or building trust with clients, knowing exactly where your data travels is essential.

Data Processing Comparison

Aspect	Cloud AI	On-Device AI
Processing location	Remote data centers	Your own hardware
Internet needed	Always required	Only for updates or optional cloud features
Data exposure	Third-party servers	Your control
Response speed	Depends on connection	Instant, typically faster
Hardware requirements	Minimal on your end	Needs capable device
Best for	Heavy processing, large datasets	Sensitive data, offline scenarios

The Bigger Picture: AI Becomes a Utility, Not a Risk

This partnership between MacPaw and Liquid AI signals a broader industry trend. As more companies follow this path, AI will increasingly become like electricity: available through your local outlet rather than requiring a connection to a distant power plant. This shift matters for Malaysian SMEs because it removes one of the main barriers to AI adoption—the fear of losing control over your data.

In the long run, we’ll likely see AI capabilities built directly into the software and devices SMEs already use. Your accounting software, your CRM system, and your inventory management platform will include AI features that run locally by default. This won’t just improve efficiency; it will change the trust calculation. When you use AI on your own device, you’re not outsourcing your intelligence—you’re simply upgrading your own.

For Malaysian business owners, the message is clear: pay attention to where your data goes today, because the tools that keep your data local are coming. When they arrive, they’ll let you harness the power of AI without compromising the trust your customers place in you. That’s an advantage worth preparing for. For more details on this industry development, you can read the original TechCrunch coverage.

Ready to Streamline Your Operations?

Your business should run itself. AutoRunBiz deploys AI agents to automate your daily operations — WhatsApp orders, invoicing, customer follow-ups, and accounting. Book a free 15-min ops audit to see where automation fits your business →