Liberty
Short-term Insurance Chatbot
Liberty
Short-term Insurance Chatbot
Liberty
Short-term Insurance Chatbot
Improve the experience of obtaining an insurance quote
Liberty wanted to explore an alternative approach to providing customers with quotes and policy management. They opted to try a chatbot approach similar to Lemonade insurance
Research Kick-Off
As a starting point we conducted creative internal workshops to get a feel for the characteristics people tend to associate with feelings of trust, professionalism and helpfulness.
What should the chatbot's personality be?
Research objectives
1. Which avatar look would be considered the most trustworthy, professional and friendly to users.
2. Which tone and voice the chat bot should employ.
3. Assess the usability of main user journeys using qualitative, moderated in-person assessments.
Challenges
1. How to simplify long, anxious user journeys.
2. Figuring out the most user-centric, chatbot personality and tone for the insurance context
3. How to develop a device-agnostic design - React Native meant finding an overall interaction model that would be usable for both iOS and Android users
4. How to create an omni-channel experience across web, mobile & Facebook
Avatar Testing Round 1
Based off the initial workshops we mocked up some potential avatars which were in line with our preliminary findings and tested these with external participants.
Avatar Testing Round 2
The avatars were then further refined and re-tested with external participants to gauge their sentiment towards each option until a final avatar was identified.
Tone and Voice
We simultaneously tested three distinct tones of voice, 'professional', 'humourous' and 'familiar' to ascertain what tone of voice would best convey a sense of trust, professionalism and helpfulness.
Participants were asked to navigate through three different conversation flows and asked questions in an interview-style while they navigated the different flows to assess their sentiment towards each flow.
Usability Testing
Multiple rounds of moderated, in-person usability tests were done using prototypes of specific journeys.
Qualitative assessment of psychophysiological and neurophysiological responses to the prototype UI + UX during the completion of specific tasks.
Retrospective probing during a post-exposure user experience interview. The qualitative responses of users are to be captured and compared with the quantitative biometric data (eye-tracking, heart rate monitoring, facial expression monitoring) for additional insight into their perceived user experience and design/interface.
Reflections
This was one of the first projects I worked on in my UX design career. Most research was done in collaboration with an external UX research company. Their usability tests were in-person and their setup tracked eye movement and measured participant's heart rates, amongst other things. The diversity of data obtained was interesting but a lot of it, ungrounded and open to interpretation. On the other hand, the sterility of the setup made participants nervous, the sessions felt more like a polygraph test, than an environment where participants felt comfortable to engage with the app naturally
Improve the experience of obtaining an insurance quote
Liberty wanted to explore an alternative approach to providing customers with quotes and policy management. They opted to try a chatbot approach similar to Lemonade insurance
Research Kick-Off
As a starting point we conducted creative internal workshops to get a feel for the characteristics people tend to associate with feelings of trust, professionalism and helpfulness.
What should the chatbot's personality be?
Research objectives
1. Which avatar look would be considered the most trustworthy, professional and friendly to users.
2. Which tone and voice the chat bot should employ.
3. Assess the usability of main user journeys using qualitative, moderated in-person assessments.
Challenges
1. How to simplify long, anxious user journeys.
2. Figuring out the most user-centric, chatbot personality and tone for the insurance context
3. How to develop a device-agnostic design - React Native meant finding an overall interaction model that would be usable for both iOS and Android users
4. How to create an omni-channel experience across web, mobile & Facebook
Avatar Testing Round 1
Based off the initial workshops we mocked up some potential avatars which were in line with our preliminary findings and tested these with external participants.
Avatar Testing Round 2
The avatars were then further refined and re-tested with external participants to gauge their sentiment towards each option until a final avatar was identified.
Tone and Voice
We simultaneously tested three distinct tones of voice, 'professional', 'humourous' and 'familiar' to ascertain what tone of voice would best convey a sense of trust, professionalism and helpfulness.
Participants were asked to navigate through three different conversation flows and asked questions in an interview-style while they navigated the different flows to assess their sentiment towards each flow.
Usability Testing
Multiple rounds of moderated, in-person usability tests were done using prototypes of specific journeys.
Qualitative assessment of psychophysiological and neurophysiological responses to the prototype UI + UX during the completion of specific tasks.
Retrospective probing during a post-exposure user experience interview. The qualitative responses of users are to be captured and compared with the quantitative biometric data (eye-tracking, heart rate monitoring, facial expression monitoring) for additional insight into their perceived user experience and design/interface.
Reflections
This was one of the first projects I worked on in my UX design career. Most research was done in collaboration with an external UX research company. Their usability tests were in-person and their setup tracked eye movement and measured participant's heart rates, amongst other things. The diversity of data obtained was interesting but a lot of it, ungrounded and open to interpretation. On the other hand, the sterility of the setup made participants nervous, the sessions felt more like a polygraph test, than an environment where participants felt comfortable to engage with the app naturally
Improve the experience of obtaining an insurance quote
Liberty wanted to explore an alternative approach to providing customers with quotes and policy management. They opted to try a chatbot approach similar to Lemonade insurance.
Research Kick-Off
As a starting point we conducted creative internal workshops to get a feel for the characteristics people tend to associate with feelings of trust, professionalism and helpfulness.
What should the chatbot's personality be?
Research objectives
1. Which avatar look would be considered the most trustworthy, professional and friendly to users.
2. Which tone and voice the chat bot should employ.
3. Assess the usability of main user journeys using qualitative, moderated in-person assessments.
Challenges
1. How to simplify long, anxious user journeys.
2. Figuring out the most user-centric, chatbot personality and tone for the insurance context.
3. How to develop a device-agnostic design - React Native meant finding an overall interaction model that would be usable for both iOS and Android users.
4. How to create an omni-channel experience across web, mobile & Facebook.
Avatar Testing Round 1
Based off the initial workshops we mocked up some potential avatars which were in line with our preliminary findings and tested these with external participants.
Avatar Testing Round 2
The avatars were then further refined and re-tested with external participants to gauge their sentiment towards each option until a final avatar was identified.
Tone and Voice
We simultaneously tested three distinct tones of voice, 'professional', 'humourous' and 'familiar' to ascertain what tone of voice would best convey a sense of trust, professionalism and helpfulness.
Participants were asked to navigate through three different conversation flows and asked questions in an interview-style while they navigated the different flows to assess their sentiment towards each flow.
Usability Testing
Multiple rounds of moderated, in-person usability tests were done using prototypes of specific journeys.
Qualitative assessment of psychophysiological and neurophysiological responses to the prototype UI + UX during the completion of specific tasks.
Retrospective probing during a post-exposure user experience interview. The qualitative responses of users are to be captured and compared with the quantitative biometric data (eye-tracking, heart rate monitoring, facial expression monitoring) for additional insight into their perceived user experience and design/interface.
Reflections
This was one of the first projects I worked on in my UX design career. Most research was done in collaboration with an external UX research company. Their usability tests were in-person and their setup tracked eye movement and measured participant's heart rates, amongst other things. The diversity of data obtained was interesting but a lot of it, ungrounded and open to interpretation. On the other hand, the sterility of the setup made participants nervous, the sessions felt more like a polygraph test, than an environment where participants felt comfortable to engage with the app naturally.