Cindy Rogers

Voice of health
Exploring voice technology
Making it easier for our customers to find information about their health insurance plan

Problem to solve
How can we make it easier for our customers to find out what they're covered for in their health insurance plan when they most need it? While also Reducing the number of calls to the helpdesk on general benefit queries.
During times of emergency or stress, customers want to easily access their health insurance information without having to find documents and read through the content to see what they're covered for. Could voice technology help ease this burden?
This project came into the Irish Life innovation hub from Irish Life Health so that we could explore and test this concept. We used mid-fidelity prototypes and conducted user-testing remotely as we were in the middle of a pandemic.
Our prototype consisted of recording answers into Alexa skills to use during the test. We had a scenario and script for the participant. The participants task was to ask Alexa various set questions and our mockup version of Alexa would ask the user some details in order to tell them about their plan benefits.





It makes it easy to access information that is normally hard to find in leaflets and documents.
User test on Alexa
Respondent.io participant
Learnings
We learned that using voice technology in this way would be more beneficial for users with general questions about their health insurance plan.
Questions like:
-
What benefits am I covered for
-
When is my renewal date up
-
Things I’m covered for in different hospitals
-
Am I covered questions for consultants
The project has since gone back to Irish Life Health for further exploration.
More on voice
We also explored a digital assistant where the user could choose to talk or type a question they have about their health plan benefits into a webpage AI portal.
We used a mid-fidelity prototype to test various conversations with our digital assistant. We added questions like 'Am I covered for GP visits?' and 'What benefits are included on my plan?'
Prototype






Utilising
behavioural science
We explored behavioural science by using a behavioural map to document friction in the voice journey. We discovered some cognitive overload with the answers our robo adviser would give. Our participants found it difficult to absorb the information our adviser was feeding back to them. Therefore, we had to re-evaluate the language used in the chat.

I find Alexa never has any problems understanding me. Their voice tech always works. This doesn't! It's needs a lot of work.
User test on Chatbot using voice
Respondent.io participant
LESSONS LEARNED
Using voice for a chatbot proved to be challenging for our participants as many have an anchor based on how Alexa and Google's voice sound.
It also didn't feel natural for users to want to press a button to activate record mode and then talk to a chatbot. Some users felt this kind of technology would be more suited to someone with accessibility needs.
These tests also helped us to gather insights into how users would phrase a question. Giving us more data for further exploration on Alexa.