Industry Engagements

I pursued my Ph.D. partly because I wanted my research to stay connected to the problems organizations are actually facing. Many of my research questions come from those problems. And industry talks and collaborations a primary way for me to share what I am learning, challenge my own assumptions, and identify new questions worth studying.

August 2026 @ HIMSS APAC

Late 2000s: Attended HIMSS as a practitioner.

2016: Attended HIMSS as a PhD student to interview entrepreneurs, investors, health systems folks, and customers for my study.

2026: Attended #HIMSS26APAC as a faculty member to share insights from that study + other related research.

A full-circle moment, indeed! And I couldn't have asked for a better moderator (Anne Snowdon). The conversation was candid and fun, covering a lot of ground and taking a few unexpected turns. We talked about why health systems still need startups, why it remains so difficult for the two to work together, how some have managed to overcome those challenges, and what all of this might mean for emerging technologies, such as agentic AI.

A few takeaways:

• A pilot is not a path to commercialization. It can help you learn more about the problem, workflow, economics, and, for start-ups, whether you even have the right customer/market.

• Health systems can help reduce some of the guesswork for startups. Startups still have to do the discovery and solutioning work. But much of what they need to know is tacit, scattered across different parts of the healthcare organization, or only becomes clear once people start working together. More concrete problem statements (e.g., via reverse pitches/ in calls for innovation) and structured access to the people who actually live the workflow can help.

• For startups, look beyond the obvious customer. If one hospital doesn't buy, moving to another hospital may not actually change the underlying economics. Usefulness ≠ willingness to pay. A more useful question is: who really has this problem, who benefits when it is solved, who carries the cost or risk, and who has a real reason to pay? The answer may not always be the customer you started with.

If you are interested, Douglas Hannah and I have a Harvard Business Review article and a full paper that discusses some of these insights.

Overall, great to catch up with old friends in the industry and meet new ones. Note to self: time flies, so remember to take more photos with friends next time!

WhatsApp Image 2026-10-24 at 4.26.24 PM.jpeg

Jan 2025 @ ASME AI Festival

Pilot Tested: Key Insights from SMEs Using Generative AI

Key takeaways based on my research on SMEs' GenAI implementation:

Finding the right use case requires experimentation. SMEs need to learn where the technology works, where it fails, and whether a seemingly good AI application actually fits how their particular industry creates value.
AI is not plug-and-play. Successful adoption requires customization, prompting, training data, testing, and repeated refinement. Firms that expect to simply install a tool and obtain value are much more likely to struggle.
The SME is always going to be the domain expert. Employees who understand the customers, workflows, and pain points are often the people best positioned to identify useful AI applications and teach the system what “good” looks like.
Data readiness matters. AI cannot compensate for missing, poorly organized, or inappropriate business data. Some SMEs abandoned projects because they lacked the customer, process, or performance data required to make the application useful.
Implementation is a joint learning process between the SME and the AI vendor. The more successful cases involved SMEs explaining their goals, processes, and constraints while vendors helped configure the solution, prepare data, and refine it through testing.
Starting small has value even when the first project fails. Early experiments build the firm's ability to evaluate AI, understand what to ask vendors for, and make better adoption decisions as the technology develops.