This person is building a data pipeline with string and chewing gum, and that is exactly the problem. A clinician with Python and R experience is spending their evenings writing brittle Selenium scripts to click buttons on a medical charting system, just to get their own patient data into a format they can actually use. This is not a skill issue. This is a system failure, and it is happening in thousands of clinics right now.
The user's workflow is instructive. They need to log into a cloud-based medical record system, navigate to specific pages, click the right elements, download reports, rename files, open them in Excel, transform the data, and finally combine everything into a single sheet. Every step is a fragile point of failure. The login script is "tough and finicky." The report names change. The whole process depends on manual folder management and hope. This person is doing what every competent spreadsheet user eventually does: they are building workarounds because the tools they have were not designed for the work they actually do.
Here is what this reveals. The medical charting system generates data, but it does not make that data accessible. The user has to fight the interface to get it out. Then they have to fight Excel to get it organized. The gap between having data and being able to use it is enormous, and it is filled entirely by user effort. A truly intelligent spreadsheet would not require this. It would connect to the source, authenticate securely, pull the reports, handle the naming variations, and present the combined results in a structure the user defines. The user should be spending their time on analysis and patient care, not on writing scripts to simulate mouse clicks.
Our take is direct: the future of data management is not about users becoming better at automating broken workflows. It is about tools that make the workflow unnecessary. This user has already identified the right approach, automation, scripting, thoughtful filtering, but they are working against their tools, not with them. An AI-native spreadsheet should be the place where that work happens, not a separate skill to learn on top of everything else. The next step is not to learn more Python. It is to demand a tool that understands the problem from the start.