To know me in real life is to know that I have a hard time sitting still. If I see a dirty dish in the sink, I wash it. If my lawn needs mowing, I’m out there with the mower. My idea of relaxation is a 40-mile bike ride.
This bleeds into my programming life where I have a very hard time calling anything finished. No matter how much work I put into a software project, there are always new ideas to explore, enhancements to add, and bugs to be fixed.
Last week, I wrote about my proof-of-concept bot for hospital patients. I also presented a video that showed my smart speaker bot in action. The bot enabled patients report and illness and schedule a nurse to stop by. The scheduling part was done with a ServiceNow workflow application invoked by the bot.
After completing the fundamentals of the bot, I realized that I had only scratched the surface of what a bot of this sort could do. In addition to requesting a nurse, why not order meals, answer FAQ-type questions, and purchase items needed during the patient’s hospital stay.
So, it was back to coding and after a few evening and weekend hours, I came up with a new and improved bot that greatly expanded on last week’s incarnation. Specifically, in addition to reporting an ailment, my new bot provides:
- An FAQ service for general patient questions.
- A chit-chat service to humanize the bot.
- A patient ordering system. For my proof-of-concept I decided that patients should be able to order socks. To order socks, the bot must know the number of pairs and the requested color. The order will be put into the ServiceNow backend application.
- A meal ordering service. The services allows that patient to order breakfast, lunch, dinner, or a snack.
In the video, I want you to notice a few things. First, I begin with the chit-chat capabilities. I ask all sorts of random questions and the bot responds appropriately. Next, I ask for a nurse. I then go into a series of socks requests. Notice how I can ask for them in all sorts of different ways. Finally, I order a meal and ask the bot to tell me a joke.
Here are the patient-bot interactions that resulted in new workflows. Notice the request for a nurse, the different requests for socks, and the meal order. In the real world, these workflows would schedule actions to take place.
Mischief Managed
I am pretty sure that I am not finished with my bot, but this is enough to show you how a bedside bot can greatly improve a patient’s hospital stay. As I said in my previous article, the sky is the limit in how this technology can be applied. This is only the beginning.
Tao, Zen, and Tomorrow 