HashMicro is a leading Enterprise Resource Planning (ERP) software provider in Southeast Asia, known for helping thousands of businesses transform their operations. Founded in 2015 in Singapore, the company also has a strong presence across Indonesia, Malaysia, and the Philippines.

As AI capabilities expand, HashMicro continues to build on that foundation, enhancing the way users interact with data, streamlining processes, and improving responsiveness to business needs in real time.

With most customers in Indonesia, local hosting was key to reducing latency. Therefore, it needed a cloud partner with a strong presence in the Southeast Asian market. HashMicro also needed a scalable cloud platform to handle fluctuating user loads, given its unlimited-user license model.

HashMicro also wanted to explore Large Language Models (LLM) to elevate how users interact with ERP data. They were looking for a model that processes natural language and generates contextual responses, adding a new layer of accessibility to enable conversational interactions through platforms like WhatsApp or internal chat, without disrupting existing workflows or logic.

Cost efficiency and responsive support were also crucial in ensuring seamless and reliable service delivery.

With data centers across Southeast Asia, Alibaba Cloud presented the perfect choice for HashMicro, providing strong reliability and reduced latency for its customer applications.

HashMicro chose Alibaba Cloud Qwen as its LLM due to its native integration, enterprise-grade architecture, and seamless alignment with its modular, API-first ERP ecosystem.

Qwen’s Vision-Language (VL) capabilities extend HashMicro’s existing OCR technology stack by offering multimodal understanding of documents. VL models can also interpret hierarchical layout, tabular structures, and embedded metadata, essential for automating invoice matching, contract versioning, and form classification within HashMicro’s Document Management System. It also enables dynamic prompt chaining and agentic workflows, allowing internal tools to summarize, categorize, and make routing decisions based on business logic embedded in the company’s finance, HR, and procurement modules.

Backed by Alibaba Cloud’s robust SLAs and support, Qwen ensures high availability, stability, and faster response times for HashMicro’s enterprise clients.

HashMicro needed a reliable compute platform to host its applications, and it achieved this successfully by using Alibaba Cloud Elastic Compute Service (ECS) instances. HashMicro benefited greatly from Gen 7 instances owing to the X-Dragon chipset, which further increased the performance of the instances. With autoscaling, HashMicro also ensured increased reliability of its solution when traffic increases and saves costs when the utilization is low.

HashMicro is built on a modular, cloud-native architecture with open REST APIs, real-time data pipelines, and flexible integration layers. This foundation allowed the company to embed Qwen smoothly into its environment. Instead of replacing existing automation, Qwen enhances it, providing users with a more intuitive way to retrieve insights, follow up on tasks, and respond to business needs in real time. It’s a natural extension of HashMicro’s long-standing focus on practical, usable intelligence.

Additionally, a highly available Server Load Balancer (SLB) was also implemented on top of the instances. The SLB has a backup instance located in a different region, which ensures the availability of HashMicro's applications in the event of a regional data center disaster. Furthermore, HashMicro also deployed a database with a standby server for disaster recovery. This setup is aimed at making the service highly available as well as redundant, reducing the possibility of downtime.

Since integrating Qwen LLM, HashMicro has significantly enhanced the intelligence layer across its ERP ecosystem, particularly in managing complex workflows, interpreting data, and improving user interactions.

● One of the most impactful advancements has been the introduction of context-aware automation across multi-step approval processes, a capability traditionally reserved for highly complex enterprise systems. The LLM now interprets business rules, user roles, and historical context to route tasks more accurately and dynamically.

● In modules such as Accounting, Procurement, and Contract Management, Qwen’s Vision-Language capabilities have allowed the company to go beyond standard OCR. The system can now read, structure, and summarize semi-structured business documents, such as purchase requests, vendor contracts, and delivery notes. This helps to autofill ERP fields, match data across systems, and reduce human validation efforts by over 40%.

● From an architectural standpoint, Qwen integrates seamlessly with HashMicros’s cloud-native, multi-tenant ERP platform, maintaining real-time performance even during concurrent processing across customer environments. It complements existing features, such as predictive analytics, event-driven automation, and modular process orchestration, ensuring the platform remains both powerful and accessible.

Looking ahead, HashMicro aims to accelerate its digital transformation by leveraging Alibaba Cloud’s cutting-edge AI and cloud technologies to build smarter, more adaptive ERP solutions. The company plans to expand its use of large language models and vision-language capabilities to automate increasingly complex business processes. HashMicro also intends to deepen its collaboration with Alibaba Cloud to enhance the scalability and security of its platform.