BlueCard Chat
SUMMARY
This decision will improve operational efficiency for the BCBSA and its 34 subsidiaries while enhancing healthcare services for millions of Americans.
As framing a brand new experience, we were facing key challenges as following:
- Picture a new communication style or workflow no one has imagined, or experienced(So how?)
- Scope down the new experience into minimum viable product(MVP), and build it brick-by-brick
As the sole designer on this project, I need to both find effective solutions and clearly communicate every steps in my process to my team. (This is the key ingredient of 'trust' soup!)
KICKOFF
Rafting through our own path
An interesting conversation that happened between BCBSA and me at an early stage of the project reflects the challenge we are facing:
"We want to build a live chat tool to help our companies communicate on claim!"
"Can you help me to elaborate on what chat functions you are imagining right now?"
"A LIVE CHAT tool! Did you get it?"
With the chat tool's scope still unclear, I took the initiative to guide our product team in defining an achievable MVP that could go from research to development within 5 months.
PROCESS
Understanding chat holistically
Since there’s a huge leap between existing email communication and envisioned chat tools, I asked myself a question before diving into the research:
"What role that our chat is going to play during user’s workflow?"
I want to take a step back to understand how the product will support users' workflows. This helps me focus on the key product offerings rather than getting lost in minor UX details.
Identifying scope with team
After conducting 12 user interviews and analyzing insights, I organized a workshop with the product team and tech leads to define the MVP scope. This workshop provides the space to sharing research process and findings, also opens for collaborative discussion. Together, we evaluated features to enhance, keep, or remove.
PROBLEM REVEALING
Ad hoc conversation in the workflow
Our JTBD exercise and user journey revealed our users work pattern: The need for conversations can pop up at any time during a day, and ad hoc attention and responses are needed.
...vs. complicated messaging tool
However, the current communication system cannot effectively handle spontaneous conversations or manage users' varying attention needs. The pain points visualized below highlight key opportunities for improving our new chat experience.
NORTHSTAR
How might we help BCBSA employees to jump into the conversation at any place, at any time to speed up their workflow?
SOLUTION
Introducing BlueCard Chat
After numerous sketches, reviews, iterations and tests, we became confident to introduce BlueCard Chat to our users.
Multitasking between pages and tabs
Integrate chat into the workflow
USER SCENARIO
Users had to open multiple windows/tabs for data transcribing or spelling check between the internal claim data and message.
SOLUTION
We broke down the page content into columns based on their usage. Users can easily multitask while chatting.
Multitasking between pages and tabs
One chat at a time
USER SCENARIO
Users could easily overwhelmed by multiple messages, resulting in a slower response.
SOLUTION
By displaying chats in card style and allowing users to continue one conversation at one time, we kept users' attention to every update of the current task. At the same time, we limited users infinite scrolling to prevent unnecessary distractions caused by other information.
Seeking Context for Conversation
Navigate users to ask right questions
USER SCENARIO
Chats can be different one and another. Currently, users spent a great amount of time on communicating context instead of solving issues.
SOLUTION
We introduced automation and categorize chats by users' questions before connecting to real expert. To help users further identify their questions, we provided the tooltip displaying the common questions under each category to choose the correct question area for users' reference.
Massive Manual Work
Copy data in one click
USER SCENARIO
Users used to use multiple clicks and countless Ctrl + C/Ctrl + V to transcribe key data value between chat and other data fields.
SOLUTION
By providing duplicate button for key data in message drawer, we simplified the copy/paste by one click for user transcribing date flexibly.
Seeking Context & Novice Experience
Helping users easily grasp Chat
Since it would be the first time we introduce a new chat functionality to users, we had to provide a soft start on user onboarding. Thus, we polished UX copy for all the empty states and loading animation. With clear direction on how users can use chat and what they can achieve via the tool, users can have overarching understanding on the usage and what they can expect in each step of chat.
NEXT STEP
A designer's job never ends
Before launching the chat tool to broad user groups and pilot plans, we had several product trials with claim processors and managers to test if our design met their working mission, "handling the claim/messaging responsively and concisely." We focused on Average Chat Handle Time as one of our key metrics and documented time delays caused by UI misunderstandings or interaction errors. We are proud of the trial receiving extremely positive feedback:
The average task completion time decreased by 15%
LEARNINGS
- Unpack the challenge from product thinking instead of user-centered thinking
During the project, I was constantly practicing 'Jobs-to-be-done' and unpacked what the user wants to achieve during their work and how our product help users reach there. We previously over-focused on user experience and pain points, leading us to spend too much time jumping into various dazzling innovations which deviate from the core value of chat or inventing UI patterns that didn't fit into users' workflow. - Best UX ≠ Happiest Path.
When I started to work on an actual project, I realized I must take care of all the edge cases and every aspect of user experience, and I had to get rid of 'only thinking of the happy path.' A smooth, happy path is only the part of UX.
Every time I finish the low fidelity user flows, I will go by each user actions/system response to see what extreme scenarios(i.e., too long user inputs)/errors(i.e.no relevant search results) could happen. Based on edge cases I found, I can reframe my flow to navigate users back to the right track.
Date
Position
Responsibilities
More
May 2022 - Aug 2021
Product Designer
User Research
User Flows
Wireframing
UX Writing
-