[Edaily Reporter Yun Junghoon ]HANWHA SOLUTIONS(009830)is accelerating the enterprise-wide adoption of generative artificial intelligence (AI) through collaborations with Amazon Web Services (AWS) and Anthropic. The company’s strategy is to proactively establish security and governance frameworks to ensure both rapid AI adoption and the protection of corporate data, while moving beyond individual productivity improvements to generate tangible business results.

Oh Sarang, Head of the AI Team at HANWHA SOLUTIONS, introduced HANWHA SOLUTIONS’ AI Transformation (AX) case study during the “AWS AI Wave” keynote speech held on the 23rd at Parnassus Tower in Gangnam-gu, Seoul, and stated, “When we decided to open ‘Claude Cowork,’ based on Amazon Bedrock, to all employees, it was difficult to find reference cases—both domestically and globally,” she said. “Because we established security and governance first, we were able to apply AI to actual work even before the general availability (GA) release.”

HANWHA SOLUTIONS, which operates global energy and materials businesses centered on solar power and petrochemicals, has set the core goal of its AI Transformation (AX) to go beyond merely allowing individual employees to use AI autonomously and instead embed it into company-wide business processes.

In the early stages, given the nature of the manufacturing industry—where even a single design parameter could constitute a trade secret—and the fact that each subsidiary had its own certification system, the adoption of AI was accompanied by challenges related to security and control. In particular, in the environment of a large corporation where most employees are not developers, establishing AI in day-to-day operations required a systematic approach that went beyond simply deploying tools.

Accordingly, HANWHA SOLUTIONS first established a common security and management framework for AI usage to mitigate risks.

The key was designing an architecture where all AI calls pass through an “LLM Gateway” within the company’s own AWS account before connecting to Amazon Bedrock. This allows the company to manage usage data and logs within its internal account and monitor usage status by subsidiary and organization in real time via a unified management console.

Amazon Bedrock Guardrails were also applied to all calls. The system is designed to automatically mask personal information—such as names and phone numbers—before passing it to the model, thereby reducing the risk of sensitive data exposure. By consolidating AI usage pathways and establishing a system for data protection and usage monitoring, the company has secured the foundation necessary for company-wide adoption.

These preparations led to faster adoption of new technologies. As soon as the Claude Cowork Research Preview was released on Amazon Bedrock last April, HANWHA SOLUTIONS began evaluating it and established a production deployment environment within three weeks. By the time of the official launch in July, the company had completed its initial proof of concept (PoC) and reached the stage where the technology could be utilized in actual business operations.

Collaboration with AWS and Anthropic also played a crucial role in the adoption process. HANWHA SOLUTIONS worked with both companies to swiftly resolve on-site issues and refine the service implementation, expanding the deployment scope from a small, dedicated team to global business sites, including those in Korea and the United States.

HANWHA SOLUTIONS’ governance model is also spreading within the group. The number of group companies currently pursuing the adoption of Claude has increased to seven. This means that the security and operational framework established by HANWHA SOLUTIONS has become a benchmark for supporting AI utilization at the group level. Global companies are also showing interest in this process of establishing enterprise-wide AI adoption.

HANWHA SOLUTIONS is now pursuing a second-phase strategy focused on creating tangible business value, moving beyond simply expanding AI adoption. While the initial proof of concept (PoC) focused on improving individual productivity and office automation, the company plans to operate a team-level budget pool to identify high-complexity challenges across core businesses and the entire value chain.

In particular, the company plans to internalize work-oriented AI agents linked to team Key Performance Indicators (KPIs) to secure use cases that translate into actual business results. The goal is not merely to increase the number of AI implementations or usage volume, but to specifically verify the impact on operational efficiency and business competitiveness. The key lies in promoting business-led AX by linking team-level goals with AI utilization.

Team Leader Oh stated, “Governance is not an obstacle that slows down AI adoption; rather, it is a decisive force that safely embeds AI into the company’s core business.” He added, “While adoption is a technical challenge, embedding it is a management challenge; true business value will only be created when AI is connected to the company’s core business.”