Hanmi Pharmaceutical, one of South Korea’s largest R&D-driven drug developers, has equipped its entire field sales force with 5G-connected Surface Copilot+ PCs and deployed Microsoft 365 Copilot across the organization. The move marks one of the first large-scale, security-anchored AI PC rollouts in the tightly regulated pharmaceutical sector—and it signals a broader enterprise shift toward co-engineered hardware-software AI ecosystems.

The deployment, detailed in a Microsoft customer story published earlier this year, comes after a proof-of-concept phase in 2024. By early 2025, Hanmi had expanded the program, combining Microsoft 365 Copilot with Surface Copilot+ PCs to create what it calls an “AI PC environment.” The company’s three explicit priorities: productivity, mobility, and security.

What Hanmi Implemented

5G-Enabled Surface Copilot+ PCs in the Field

Sales representatives now carry Surface Copilot+ PCs with built-in cellular 5G. Previously, field workers used tablets but often had to return to the office to access internal systems due to security constraints, creating delays and lost sales opportunities. The new devices, running on Qualcomm Snapdragon X Elite or Plus chips with dedicated Neural Processing Units (NPUs), allow reps to stay connected anywhere.

“All sales personnel in the field have been equipped with 5G-enabled Surface Copilot+ PCs, enabling a connected digital work environment anytime, anywhere,” said BumJin Kim, Group Leader of the Information Strategy Group at Hanmi Science. “Using the Surface Copilot Key, employees can quickly and intuitively search for the information they need while accessing and managing cloud-based storage like OneDrive and SharePoint from outside the office.”

Key device-level capabilities:
- Persistent, low-latency 5G connectivity for cloud access beyond Wi-Fi coverage.
- On-device NPUs that accelerate local AI tasks, reducing reliance on cloud round-trips for everyday queries.
- A dedicated Surface Copilot Key—a hardware button that instantly invokes Copilot for natural-language search and contextual assistance.
- Enterprise-grade security stack, including Pluton hardware root of trust, integrated with Microsoft Intune and Sentinel.

Microsoft 365 Copilot Across Workflows

Hanmi wove M365 Copilot into daily operations in Word, Excel, Outlook, and Teams. The sales planning team reported immediate gains.

“In the past, we managed performance and handled work requests via email or messaging, but now we save forms on SharePoint and collaborate in real time, saving time and effort,” said YongBae Kim from the Sales Planning Team. “We are leveraging the town hall and breakout room features on Teams to suit different work situations, enabling more efficient meetings and training sessions.”

The Data Strategy Group, which handles marketing and business development, found particular value in AI-powered meeting intelligence. Naeun Kim from that team said: “We frequently hold meetings with external partners, and features such as scheduling, automated meeting summaries, and documentation have greatly reduced the steps required for communication with partners.”

These productivity bursts align with broader industry reports showing that Copilot users can save hours per week on writing, data analysis, and meeting documentation—a critical edge in a competitive pharma market.

A Zero Trust Security Foundation

Hanmi, already ISO 27001 certified, layered Intune for device management and Microsoft Sentinel for SIEM to enforce centralized control and real-time threat detection across its new fleet. The combination lets the IT team:
- Track device location and connectivity, enabling remote wipe if a device is lost.
- Enforce conditional access policies and centralize app distribution via Microsoft Entra ID.
- Aggregate telemetry and alerts for 24/7 threat hunting.

This architecture maps cleanly to Zero Trust principles, a necessity when field devices routinely handle sensitive clinical and commercial data outside the corporate network.

Business Impact: From Faster Collateral to Citizen Developers

The most immediate win is sales mobility. Reps no longer need to return to the office to update records, find product collateral, or submit forms. The Surface Copilot Key and on-device search mean they can pull up the latest clinical data or pricing charts in seconds during a customer visit, reducing administrative drag and creating more face-time with physicians.

But Hanmi’s ambitions extend beyond the sales force. The company is actively training “citizen developers” through Power Platform workshops and an early-adopter program. The goal: enable business teams to build their own automations (Power Automate flows, Copilot Studio agents) without overwhelming central IT. This bottom-up strategy accelerates process digitization and creates a workforce that can adapt AI tools to domain-specific needs—a model many enterprises will watch closely.

Technical Architecture: Hybrid AI and Hardware-Rooted Security

The Surface Copilot+ PC line exemplifies a hybrid AI model that balances local inference and cloud compute. Lightweight tasks—like contextual searches, real-time translations, or camera effects—run on the NPU using small language models (SLMs) stored on-device. Heavier workloads, such as cross-document summarization or complex reasoning, are offloaded to Azure-hosted large language models (LLMs). This split reduces latency and bandwidth consumption while keeping sensitive data local when possible.

Independent reviews of Copilot+ PCs confirm that this architecture yields tangible responsiveness gains, especially for repetitive AI-assisted tasks. However, the full benefit requires apps optimized for the Windows Copilot Runtime; currently, many third-party tools still rely solely on cloud inference.

Security is another architectural pillar. Surface Copilot+ PCs ship with Microsoft Pluton, a hardware security processor that protects encryption keys, credentials, and other secrets from physical attacks. Combined with Intune’s policy engine and Sentinel’s SIEM, Hanmi can enforce a uniform security posture across its entire device estate. For regulated industries, this means every AI interaction—and the data it touches—can be logged, monitored, and audited.

Operational Realities: What IT Leaders Must Plan For

As promising as Hanmi’s blueprint appears, community analysis from WindowsForum highlights several gritty details that organizations cannot afford to overlook.

5G Connectivity and Data Costs

While 5G offers always-on connectivity, real-world performance depends on carrier coverage. Rural sales territories may see spotty signal, and continuous use of data-hungry AI features (e.g., streaming Live Captions in Teams, large file syncs) can inflate mobile data bills. IT teams should use Intune to throttle non-essential background sync on cellular and monitor usage with budgeting policies.

Device Lifecycle and Support

Deploying hundreds of 5G laptops to a mobile workforce introduces provisioning, repair, and warranty complexity. Hanmi likely uses zero-touch provisioning through Windows Autopilot, but organizations must still plan for rapid device swaps, manage eSIM profiles, and negotiate carrier contracts—tasks that fall outside traditional PC refresh cycles.

Data Governance and Copilot Hallucinations

Generative AI models, even when grounded in organizational data, can produce inaccurate or fabricated outputs. For a pharmaceutical company, a hallucinated claim in a regulatory submission could have severe consequences. Hanmi’s adoption of Sentinel and Intune is a start, but the forum analysis stresses the need for strict data classification, Copilot usage policies for regulated content, and mandatory human review for any AI-generated output bound for authorities.

Audit Trails and Compliance

Every Copilot action that affects records should be traceable to the employee, the data sources, and the model version used. Hanmi’s integration of Entra ID and Sentinel provides a foundation, but IT leaders must configure detailed logging and retention policies to satisfy auditors. This is still an evolving area; Microsoft’s compliance documentation for Copilot is improving, but enterprises should pilot these controls before scaling.

A Blueprint—Not a Recipe

Hanmi’s story is both a bellwether and a cautionary case. It proves that a well-orchestrated AI PC strategy can deliver measurable productivity gains in a high-stakes industry, but it also underscores the discipline required to make it secure and auditable.

For Windows-centric IT shops, the takeaways are clear:
- Hardware choice is no longer a commodity decision. NPUs, Pluton, and 5G support have direct impact on AI latency, security posture, and field productivity.
- Copilot governance must be a first-class priority, with policies, DLP rules, and audit trails live before the first user logs in.
- Investing in citizen developer programs can multiply the value of AI tools while keeping central IT from becoming a bottleneck.

Hanmi plans to deepen its AI push further with Copilot Studio agents and expanded Power Platform usage into 2026. While those plans are subject to product roadmaps and licensing changes, the direction is unmistakable: the AI PC era isn’t about a single assistant in the taskbar—it’s about weaving local AI, cloud intelligence, and identity-aware security into the fabric of everyday work. For regulated enterprises, Hanmi’s journey offers a compelling template, and a warning that the gap between a successful AI rollout and a compliance disaster is narrower than it appears.