HashKey accelerates the implementation of its AI strategy: from organizational efficiency improvement to next-generation digital financial infrastructure.

Against the backdrop of global enterprises accelerating the application of artificial intelligence, HashKey Group, an Asian digital asset financial services group, is also systematically promoting the implementation of its AI strategy. Recently, the group officially established a "Group Technology Coordination Committee," under which the company's management directly coordinates the overall planning and implementation path of AI and cutting-edge technologies. According to an internal notice, this committee will be responsible for the top-level design of the group's technology architecture, AI strategic planning, and cross-business line technology collaboration, accelerating the organization's transformation towards a higher level of intelligence and automation.

This organizational move also marks a new stage in HashKey's exploration of AI. Previously, AI existed within the company primarily as a personal tool and for limited trials; however, with improved model capabilities and a more mature organizational understanding, the company has begun to systematically promote AI at the group level and is attempting to integrate it into internal operations, R&D processes, and user service systems. Recently, HashKey Group CTO Devin Zhang shared his observations and assessments regarding the company's AI strategy, security framework, and application prospects in the digital asset industry.

Devin believes that we have now entered a critical period for enterprises to adopt AI on a large scale. On the one hand, basic model capabilities have reached a high level, capable of supporting more enterprise-level applications; on the other hand, after several years of market education and practical experience, organizational and talent preparedness is gradually improving. Against this backdrop, HashKey's understanding of AI is closer to an upgrade of organizational capabilities across the entire process.

Q: Why is HashKey systematically advancing AI at this stage?

Devin Zhang: Over the past year or two, AI has already entered the practical application stage within the company, especially in the R&D team, where AI-assisted programming is quite common. The real change now is that AI's role is evolving from an efficiency tool at the individual level to a capability system systematically introduced at the company and group level. This is a critical juncture for two main reasons. First, the basic capabilities of large-scale models are relatively mature. Although still iterating rapidly, the underlying capabilities can already support enterprise-level applications. Second, the preparedness of the organization and talent is also gradually maturing. After several years of exposure and use, everyone has formed a basic understanding and accumulated experience. For HashKey, the goal of advancing AI is to combine human judgment with AI's execution and efficiency-enhancing capabilities to support larger-scale business growth under the premise of compliance and controllability.

Q: What is the most direct value of AI for organizations like HashKey?

Devin Zhang: I think there are two main aspects. One is the systemic improvement of internal operational efficiency, and the other is the continuous upgrading of user experience. Regarding internal operations, we are currently focusing on two main lines. One is improving the efficiency of the R&D chain, which means gradually introducing AI capabilities from requirement breakdown, design, development, testing to deployment and delivery. The other is the non-R&D chain, including teams in HR, legal, finance, compliance, marketing, and public relations. Many departments are using AI now, but it's mostly used in isolated instances. What's truly important is full-chain AI implementation, which improves the collaborative efficiency of the entire organization. From the user's perspective, many future financial service interactions will shift from operation-driven to intent-driven. Users will express what they want to do, the system will understand the intent, organize the execution path, and provide user confirmation at key points.

Q: Which AI scenarios will HashKey prioritize for implementation at this stage?

Devin Zhang: We focus more on scenarios with clearly defined business objectives, long processing times, high repetition rates, and measurable efficiency improvements. In the R&D system, this approach is mainly reflected in the AI-driven transformation of the R&D CICD process; in non-R&D departments, it is more reflected in the automation of various highly repetitive processes with clearly defined rules.

Meanwhile, HashKey has also been advancing the application of AI in infrastructure security and risk control. In infrastructure security, HashKey has used AI for threat hunting, case tracing, potential risk discovery, and IT asset management, helping to improve the depth and breadth of its overall security capabilities. In risk control, intelligent agent collaboration mechanisms have been introduced for anti-money laundering investigations, group behavior identification, and judgment of some complex account issues. Some support processes that previously required lengthy processing by the risk control team can now be analyzed by intelligent agents first, and then reviewed and delivered by relevant personnel. On the R&D side, HashKey has also developed a clear plan for the AI-driven end-to-end development process. Related explorations are gradually advancing around requirements understanding, code generation, testing, and pre-deployment auditing, and application security auditing is also being gradually incorporated into the R&D AI pipeline.

Q: Why is a security framework a prerequisite for financial institutions to advance AI?

Devin Zhang: Because the objects of governance have changed once AI enters business processes. In the past, people focused more on the strength of the model and the accuracy of its answers. But when an AI agent truly enters enterprise processes, it faces system access, API calls, data reading, process execution, and even external action triggers. At this point, what enterprises need to manage is a set of execution entities that can access resources, invoke permissions, and complete actions. For financial institutions, the focus of risk also shifts more towards the intelligent agent architecture, permission management, and execution boundaries. To enable intelligent agents to perform tasks, permissions need to be granted to them; once permissions enter the real business flow, permission boundaries, key management, resource access rules, behavior tracking, and responsibility attribution all need to be redefined. Only after a layered, decentralized, controllable, traceable, and auditable governance system is established can AI truly enter core business processes.

Q: As a compliant exchange, how will HashKey manage the pace and boundaries of its AI development?

Devin Zhang: We will systematically advance AI, and we believe it will have a significant positive impact on the industry, the company, and our business. However, the pace of implementation needs to be aligned with the regulatory environment. At this stage, priority should be given to areas such as internal efficiency, back-end capabilities, risk management, and AI integration into the R&D process, because their value is clear, efficiency improvements are highly certain, and external risks are easier to control. Regarding user-side innovation, especially capabilities that directly enter the trading process and could significantly change users' trading frequency and behavior, the pace of implementation will be more cautious. Intent-driven interaction can help users simplify operations, but when intelligent agents further automate strategy execution or even replace users in making higher-frequency trading decisions, compliant exchanges need to more fully assess the boundaries of responsibility, user protection, and regulatory requirements. Whether such capabilities should be integrated into the product system is more suitable for gradual implementation while maintaining consistency with the regulatory environment.

At the infrastructure level, HashKey will adopt a parallel approach of using external enterprise-grade platforms and private deployments. Currently, the company will primarily utilize platforms with organizational-level data isolation, security accountability, and multi-model invocation capabilities; for more sensitive scenarios, the private deployment path will be retained. HashKey itself will focus on building business-specific intelligent agent systems on top of its basic models.

Looking at a longer timeframe, HashKey's understanding of AI extends to the evolution of digital financial infrastructure. Xiao Feng, Chairman and CEO of the group, recently mentioned in a media interview that artificial intelligence and encryption technology are gradually moving towards deep integration. With the rapid development of AI agents, future intelligent agents may gradually possess independent digital identities and payment capabilities, and assume more roles in the on-chain economic system. Under this trend, blockchain technology may also become an important infrastructure for managing and coordinating AI agents.

Devin believes that AI will gradually change how financial service institutions interact with users, their back-end capabilities, and their technological architecture. For HashKey, the short-term goal is to improve efficiency and user experience, the medium-term focus is on strengthening back-end capabilities and technological foundations, and the long-term goal is to participate in the evolution of next-generation digital financial infrastructure. For a compliant digital asset institution, this path forward is more realistic.