In just five months, a robot “born for nurses” has come to life. China Medical University (CMU), together with its subsidiary EverBot Technology, developed Eirbot, designed to quickly address hospital staffing shortages. The robot can initially handle around 30% of a nurse’s workload. After “interning” at CMU Hospital, Eirbot will not only be used in its home institution but is also planned for broader deployment.
On the 20th floor of CMU Hospital’s medical building, the evening lights shine brightly as nurses bustle about. Amid them, a roughly 150-cm-tall figure quietly moves through the corridors—not a nurse, but the latest frontline nursing robot, Eirbot.
With nurse shortages intensifying in recent years, Taiwan’s healthcare system faces unprecedented challenges. Against this backdrop, CMU took the initiative to collaborate with EverBot Technology to develop Eirbot. During an interview at CMU, the hospital’s president Chou Te-Yang and EverBot chairman Y. C. Lee shared the origin and vision behind this technological innovation.
Three Parties Unite: “Born for Nurses” as a Core Belief
In June last year, CMU Chairman Chang-Hai Tsai asked Y. C. Lee to assess the university’s robot R&D initiatives. Lee quickly realized that, given the urgent nurse shortage, it was more practical to shorten development timelines and focus on simple, functional robots rather than pursuing complex, futuristic technologies. He asked himself: “Who should this robot serve?”
Lee concluded that the most pressing need in hospitals is nurses. The positioning of Eirbot as “born for nurses” became the guiding principle. Chou explained that nurses have complex roles—“like caregivers, they handle everything and ‘protect’ everything.” Only by relieving part of their workload can nurses focus on patient care and help ease the strain on medical teams.
Through a fortuitous connection with Pan Chien-Cheng, chairman of Phison Electronics, Lee learned of Phison’s aiDAPTIV AI edge computing technology. Both parties recognized the potential for integration. Combining CMU’s clinical expertise with Phison’s AI and EverBot’s robotics, the three parties formed a seamless collaboration.
Prioritizing Practicality Over Flashy Features
Eirbot’s development was remarkably swift. In April, Phison and EverBot jointly founded EverBot Technology. By mid-September, the Eirbot prototype was ready—just five months from conception to reality, a rare feat in Taiwan’s smart medical device industry.
Lee attributes the rapid progress to the shared principle among CMU, EverBot, and Phison: practical medical application first. Chou emphasized that the development focused on core nurse needs. “Our sole goal is to save time and effort for frontline staff. Flashy features are secondary.” Every function in Eirbot was repeatedly discussed and validated by hospital nurses, integrating extensive clinical experience.
Chou also challenged the R&D team with practical scenarios. Lee recalled, “We tested dialogue examples, and the president immediately asked, ‘If the user speaks and the robot responds after 10 seconds, who will wait? And if the response is thousands of characters, the person will have walked away!’ We adjusted immediately.” Now, Eirbot responds within 3–5 seconds, keeping answers around 80 words—matching real conversation rhythm and meeting the high standards set by CMU.
Optimized Backend AI to Expand Applications
Eirbot’s ability to assist nurses relies on a powerful backend AI system. Lee explained that the robot serves as the frontend interface, while the true “brain” resides in backend servers, with a single AI server capable of connecting 20–30 robots simultaneously. Integration with the hospital information system (HIS) allows Eirbot to identify user background, identity, and medical history, providing tailored responses efficiently and securely.
Building on Eirbot’s success, EverBot plans to introduce multifunctional robots capable of cleaning, disinfection, and delivery, as well as long-term care and home service models. Lee stressed, “Robots are just a small part of the smart medical service system. The data chain behind them is ten to a hundred times larger than what we see.”
Local Servers Safeguard AI Autonomy
Lee insisted on using local servers rather than cloud-based AI like ChatGPT. “While cloud AI is fast, it essentially gives all conversation data to the provider. In healthcare, even a single error is unacceptable. Local servers prevent privacy breaches and ensure zero risk to patient data.”
Eirbot only conveys verified medical knowledge approved by CMU physicians. Lee emphasized, “All Eirbot data is reviewed by CMU; it is not allowed to search online to avoid errors.” The team’s strict commitment to medical data sovereignty and privacy sets Eirbot apart from other AI robots.
Subscription Model Ensures Continuous Improvement
Eirbot will be marketed on a subscription basis, estimated at NT$500,000–800,000 annually. Lee explained that medical knowledge evolves rapidly; a one-time purchase risks the robot becoming outdated. Subscriptions allow EverBot to continuously update and enhance backend servers, ensuring long-term performance—similar to how Apple maintains iPhone software updates.
Chou added, “Subscription also creates a competitive barrier. As our system evolves and new robot versions become smarter, competitors trying to follow the same path will struggle to catch up.”
Practical Goals: Relieve 30% of Nursing Work
The development team takes a pragmatic approach. Chou emphasized that the goal is not to replace nurses but to support them. Eirbot’s first-generation functions include environmental guidance, health education consultation, item delivery, and remote rounds. The second-generation model will add non-contact physiological monitoring and environmental surveillance, followed by medical device certification.
Lee stressed, “We focus on realistic capabilities, not flashy features. Medical robots require high reliability, privacy protection, and a professional knowledge base—these conditions must be met first.”
From Hospitals to Remote Communities: Pioneer of Smart Healthcare
Eirbot is currently undergoing a three-month POC (proof of concept) at CMU. “After December, we plan to begin mass production, initially deploying one or two units per floor, with gradual expansion,” Lee said.
Chou envisions Eirbot in remote indigenous communities, local clinics, and even supporting telemedicine via video consultations. Early demonstrations have already drawn interest from pharmacy and obstetrics departments.
Taiwanese tech companies increasingly aim to leverage local expertise to reduce nurse workload and improve patient care. CMU and EverBot’s development experience demonstrates that when hospitals actively participate in innovation rather than being passive users, smart healthcare can truly take root—offering valuable lessons for the global smart medical industry.
Resource: 醫院自己打造護理機器人！中醫大主導×長聯研發，實現「醫療主權AI」
