As hospitals across Asia accelerate investment in artificial intelligence, robotics and automation, experience from real-world deployments is increasingly showing that successful implementation depends on much more than the underlying technology.
At AIMX Singapore, during the session “Smart Hospitals in Practice: Expectations, Reality and Lessons from AI & Robotics Deployments in Singapore and the Region”, speakers highlighted a shift away from treating AI initiatives as vanity technology projects towards using digital tools more strategically to build healthcare capacity and productivity.
This transition is being driven by two converging pressures. Healthcare systems are preparing for ageing populations, workforce constraints and growing demand, while advances in generative AI and foundation models are making increasingly capable technologies easier to deploy.
However, the discussion highlighted that technical capability alone does not guarantee adoption.
A recurring lesson from healthcare deployments is that technology needs to solve a clear problem for the people expected to use it. Healthcare professionals are more likely to adopt a new system when the benefit is immediate and tangible, whether that means reducing repetitive documentation, simplifying handovers or helping staff complete work more efficiently.
This also changes how hospitals need to communicate the value of new technologies.
Rather than focusing primarily on technical capabilities, implementation teams can frame automation in terms of what it gives back to healthcare workers, such as reducing administrative workload and freeing time for direct patient interaction.
The discussion compared this to the appeal of a robot vacuum. The objective is not necessarily that the machine performs every task better than a person, but that it removes repetitive work and allows people to spend their time on activities where human involvement creates greater value.
In healthcare, this means using robots and automation to allow clinicians and other healthcare professionals to dedicate more time to patient care, communication and empathy.
The human factor also represents one of the largest components of successful technology deployment.
A 10-20-70 framework highlighted during the session suggested that approximately 10% of implementation relates to the AI or algorithm itself, 20% to technology and user interfaces, and 70% to people, workflows and change management.
This reflects a common challenge in healthcare technology projects: organisations may focus heavily on technical performance while underestimating how significantly a new system changes existing processes and staff responsibilities.
Early involvement of healthcare users was therefore identified as an important part of implementation.
Rather than introducing a completed technology into an existing workflow, hospitals can involve users from the beginning to define the problem, understand current processes and determine how those workflows will change after deployment.
This approach can also create greater ownership among users.
Staff concerns around safety and job displacement also need to be addressed early. When AI and robotic systems are introduced, two of the first questions from healthcare workers are often whether the technology is safe and whether it will replace them.
Clear communication, education and knowledge sharing around these concerns can therefore become as important as the technology itself.
The relationship between healthcare providers and technology companies may also need to evolve.
Instead of a conventional vendor-buyer model, successful implementation can involve hospitals and technology companies becoming learning partners. Solutions may change substantially from the initial proposal as both sides gain experience from deployment and better understand the operational environment.
Another important lesson is that implementation does not end when a system goes live.
Hospitals often dedicate substantial resources to preparing for deployment, but meaningful outcomes emerge only after the technology begins operating in real workflows.
Continuous improvement approaches, including Plan-Do-Study-Act cycles, can help organisations identify problems, refine workflows and improve performance following implementation.
Taken together, these lessons suggest that the next stage of smart hospital development will be less about introducing individual AI or robotic technologies and more about integrating them into healthcare systems in ways that expand capacity.
For healthcare organisations preparing for demand in 2030 and beyond, the strategic question is increasingly shifting from which AI project to launch next to how technology can allow existing workforces and infrastructure to deliver more care without requiring resources to grow at the same rate.
The experience of AI and robotics deployments in Singapore and across the region points to a relatively simple principle: successful smart hospitals are built not only around intelligent technologies, but around the people and workflows required to make them useful.
