Artificial intelligence is rapidly expanding across healthcare, but the distance between demonstrating what a technology can do and embedding it into everyday care remains significant.
Across Southeast Asia, hospitals and healthcare providers are exploring AI for patient engagement, workflow automation and operational efficiency. Yet successful adoption depends on more than the sophistication of an algorithm. Technologies also have to fit existing workflows, address clearly defined problems and provide measurable value to healthcare organisations and their patients.
Tan Wei Chun, Regional Sales Manager at Bot MD, works with hospitals, healthcare providers and technology companies across Southeast Asia to deploy AI-powered patient engagement solutions. He discusses where AI is already delivering practical value, why some healthcare technology projects struggle to scale and what providers should consider as generative and agentic AI become more widely adopted.
Where do you see the biggest gap today between the potential of AI in healthcare and what is actually being implemented across Southeast Asia?
There is a lot of excitement around what AI can potentially do, but the biggest gap today is not necessarily AI capability. It is the last mile of implementation.
Healthcare organisations are complex. There are different stakeholders, workflows, systems and priorities, so a solution may work technically without necessarily fitting naturally into how care is actually delivered.
The real question is whether the AI can work with hospital-approved information, follow the organisation’s rules, connect with existing systems, know when to hand over to a person and help the patient complete the next step.
That is why the most successful projects tend to start with the problem rather than the technology. They focus on a specific point of friction — for example, patients waiting too long for answers, staff repeatedly handling the same enquiries, or patients dropping off before making an appointment — and use AI to improve that workflow.
So I think the next phase of healthcare AI is less about proving what AI can do in a demo, and more about proving that it can work reliably and create value in the real healthcare environment.
Where is AI-powered patient engagement delivering the most practical value for hospitals and healthcare providers today?
One of the biggest areas of practical value is in reducing friction at the front door of healthcare.
Hospitals manage a huge volume of repetitive patient interactions every day — questions about services, appointments, preparation instructions, follow-ups and many other touchpoints. Individually, these may seem simple, but at scale they create a significant operational burden for healthcare teams.
At the same time, many of these interactions are more than just routine enquiries. They are often signals of patient intent. A patient asking which specialist to see, whether a service is available or how to book is usually trying to take the next step in their care journey.
AI can help hospitals respond instantly and 24/7, guide patients to the right information or service, capture appointment intent, automate reminders and follow-ups, and escalate to a human when needed.
An important part of this is also meeting patients on channels they already use, rather than asking them to learn a new system or download another application.
There is also significant value in what hospitals can learn from these conversations. Patient enquiries can reveal what patients are looking for, where they are getting confused, which services are seeing demand and where patients may be dropping off before taking the next step. AI can help turn those interactions into actionable insights for hospital teams, rather than leaving them buried across thousands of individual conversations.
So the value is not just about reducing workload. It is about helping more patients move from having a question to actually accessing care, while allowing healthcare teams to focus on the interactions that genuinely require a human.
Why do promising healthcare AI pilots often struggle to move into wider deployment, and what separates solutions that scale from those that do not?
A pilot can demonstrate that a technology works, but scaling is a different challenge.
Once you move into a broader deployment, the question is no longer just whether the AI can perform a task. It is whether the entire workflow works in practice — with the right integrations, hospital-approved information, security and governance, clear ownership, and defined rules for when a human needs to step in.
It is also important to define success from the beginning. If the organisation cannot clearly measure what the AI is improving, it becomes much harder to justify scaling it. Depending on the use case, that could be faster response times, fewer enquiries handled manually, higher appointment conversion, better follow-up completion, or reduced administrative workload.
Stakeholder alignment also matters. The people buying the technology, the teams using it and the patients interacting with it may all have different priorities.
The solutions that scale tend to be the ones that are built around those operational realities from the start, rather than expecting the healthcare organisation to change around the technology.
In other words, a successful pilot proves the AI can work. A successful deployment proves the workflow can work at scale.
How important is integrating AI into existing clinical and operational workflows rather than asking healthcare professionals to adopt entirely new ways of working?
It is extremely important.
Healthcare professionals are already very busy, so if a technology creates additional steps or requires them to significantly change the way they work, adoption becomes much harder.
The best technology should fit into the workflow and remove work rather than create more work.
That means understanding what the healthcare team is doing today, where the bottlenecks are and which parts of that process can realistically be improved.
Sometimes the biggest impact does not come from replacing an entire workflow. It comes from automating one repetitive part of it, making information easier to access or helping teams focus their attention where human judgement is actually needed.
That is also why implementation requires close collaboration with healthcare teams.
The people actually doing the work understand the operational problems better than anyone else. Technology companies need to listen to them and build around those realities.
How do healthcare needs, digital maturity and adoption challenges differ across Southeast Asian markets?
Southeast Asia is not one healthcare market.
Every country has different healthcare systems, regulations, infrastructure, procurement processes and levels of digital maturity.
Even within the same country, a large private hospital group and a smaller healthcare provider can have very different technology capabilities and priorities.
That means you cannot simply take one implementation and assume it will work everywhere else in exactly the same way.
Localisation is important, but in healthcare it has to go beyond language to the workflow itself. You need to understand how patients interact with the healthcare system, which communication channels they use, how hospitals operate and what the local regulatory and commercial environment looks like.
At the same time, there are common pressures across the region.
Healthcare providers are trying to manage growing patient demand while dealing with workforce and resource constraints. That creates a strong opportunity for technology that can genuinely improve productivity and patient engagement.
So while there is no single Southeast Asian AI playbook, the principle is consistent: start with how patients behave and how the healthcare organisation already works, then adapt the technology around those realities.
Looking ahead, how do you expect AI-powered patient engagement and healthcare delivery to evolve over the next few years?
AI will become increasingly embedded into healthcare workflows rather than sitting as a separate tool.
Traditionally, digital patient engagement has focused on answering questions such as how to book an appointment, which specialist to see or what to prepare before a procedure. Today, AI can already help patients move through those next steps — guiding them to the right service, booking an appointment, providing preparation instructions, sending reminders and supporting follow-up within the same conversation.
At Bot MD, this is already how we approach patient engagement: going beyond simple FAQs to AI agents that support more of the patient journey across channels such as WhatsApp, Messenger and Viber, while working with the hospital’s existing systems and workflows.
As AI takes on more complex tasks, integration, governance and human oversight will remain critical. The goal is not to remove healthcare professionals from the process, but to let AI handle more of the repetitive and administrative work around care while healthcare professionals focus on the areas where judgement, expertise and empathy matter most.
The future is not just AI that answers. It is AI that helps patients get things done, while giving healthcare teams more time to focus on the work that truly requires human judgement and care.
