It happens to every hospital or healthcare provider: A crisis or emergency erupts and suddenly patients are calling in droves with urgent questions — overwhelming your patient-care staff. You know this aggravates your patients and you want it fixed now.
You’ve probably heard about intelligent virtual assistants (IVAs), which use machine learning to automate a full range of activities in the patient support process. And you’re wondering: Can IVAs lighten the load on your call-center crews?
They can, but not soon enough to put today’s fire out. When you need help in time frames measured in days, not months, it’s best to think in terms of a journey to IVA, with small intermediate steps that lay the foundation for advanced automation a few months down the road.
While the initial phases do not put every fire out, they still provide priceless data about users’ habits. This data fuels intent-analysis processes that will help your IVA system make smart decisions down the road. Over time, user-intent signals will help you predict future patient support needs.
Thus, all of your patience and preparation early in the journey pay off all along the route. There’s no wasted effort.
These are three optimal phases for making the transition to IVA capability:
Phase 1: Add One Bot to Your Call Experience
It’s best to start with a single on-ramp in your journey to an automated patient experience. It could be a simple webchat bot that asks a patient one or two basic questions before forwarding the call to a human representative. This simple solution automatically eases the strain on your staff and doesn’t take long to implement.
Remember: Everything your patients do in this webchat bot will generate data that trains the IVA system to make smarter decisions as time passes. You can use any channel your patients prefer. The point is, you want to work the bugs out of the first automated process before moving on to more complex capabilities.
Phase 2. Automate Answers to Patient Questions
With your basic bot up and running, it’s time to add a few more on-ramps to your IVA journey.
That means using bots to answer more difficult questions and automate more of the patient support experience. You might also customize the responses depending on the channel — phones, email and text messaging, for instance.
Now, you have to focus more deeply on thorny issues like authentication: validating the identity of each person and remembering their favorite authentication methods.
Here, it becomes more important to train your support staff properly and confront any shortcomings in your backend IT systems. You need to make sure your legacy systems have enough power to process complex learning algorithms.
Again, you’re feeding data into your intent analysis process to ensure that each user can log in with their favorite methods in the future. Their questions and your automated answers will produce smarter bots, streamline the experience and improve patient satisfaction.
Phase 3. Implement an IVA System
Phase 1 and 2 gave you a firm foundation to implement an intelligent virtual assistant that can respond automatically to calls from a wide range of patients and tailor these responses to their unique preferences.
A full IVA system identifies callers and remembers their habits and behaviors on previous calls. Learning algorithms teach it to take on more complex challenges because it also folds in data from the rest of your patient calls.
The ultimate goal is to free up your patient-support staff to handle jobs that computers aren’t good at — like using their intuition and training to provide a savvy, humane response to a complex patient support issue. That helps improve job satisfaction, potentially reduce turnover costs.
The DMI Advantage in IVAs
DMI has an extensive track record with IVAs in multiple sectors like health care, finance and government. Our vast pool of talent in business consulting, automation, system architecture, machine learning, data science and patient experience ensures we have the right skills. Our focus on Agile methodologies gets the right solution into our clients’ hands in tight time frames. When you have big fires to put out, these are the skills you need.
–Niraj Patel, director, artificial intelligence