Help Center Chatbot Case Study
Role: Content & UX Design
The Problem
The J.P. Morgan Commercial Card help site is not a resource directly accessible to clients during their task flows. Instead, the help site has its own domain with a pop-up window for each help prompt. Instead of leveraging the help site, many customers rely on account personnel to walk them through issues they encounter during their tasks on the site.
The Solution
Consolidating the main site and help site & allowing customers to use the product with minimal personnel training
My process
I compartmentalize my greater process into three main categories. This allows me to see both a big picture of how my building will take place as well as an understanding of how each individual step contributes to another and how they connect as a whole. I generally begin my process with an empathizing phase; due to time constraints, it became preliminary first step before the Brainstorming phase and was a constant benchmark throughout the process.
01
Brainstorm
Understand problem & define direction
02
Conceptualize
Synthesize information and begin crafting solution
03
Design
Implement concept and build solution
<span data-metadata=""><span data-buffer="">Brainstorm
Assumptions | What I would have wanted to know and do
01
Product
I assumed the mobile product and web product had slightly different functionalities from one another
02
Data Archive
I would’ve wanted to analyze existing data surrounding the main commercial card site and help site
03
Pain Points
I would’ve wanted to know any pain points users faced while interacting with both the commercial card site and the help site
Chatbot
During my brainstorming session, I began to gravitate towards the chat bot solution
My process
While brainstorming, I began to notice the chatbot could encompass a variety of solutions and offer relief to multiple pain points.
Micro & Macro
Utilizes a combination of micro copy and long form content
In task help
Allows clients to access help while still in task flow
Multi-resource
Can provide a segue to various forms of additional resources
<span data-metadata=""><span data-buffer="">Research Findings
<span data-metadata=""><span data-buffer="">3x
Micro & Macro
Spending on cognitive and AI systems have had a 3x increase since 2018
24/7
24/7
Allows clients to access help while still in task flow
<span data-metadata=""><span data-buffer="">1/3
Micro & Macro
Utilizes a combination of micro copy and long form content
<span data-metadata=""><span data-buffer="">2/3
In task help
Allows clients to access help while still in task flow
Conceptualize
<span data-metadata=""><span data-buffer="">In the conceptualization phase, I focalized the following items
Consolidation
Reducing repetitive content and bringing information into a central location
Repository
How content can be governed to support this new format
Style & Product Guidelines
How content is styled to assist and resonate with the client base in a scalable way
Card Sorting
Before beginning the conversation flow, I wanted to conduct a card-sorting activity with research participants. Card-sorting would have allowed for information allocation that is more intuitive to how a client might navigate certain topics and issues
Card Sorting
As I approached the hands-on designing and building of the chatbot, I mapped out a conversation flow that would guide the content distribution and interaction of the chatbot.
Design
The design featured a brief interaction a customer might take if searching for log in info through the chatbot.
The chatbot would be onsite readily available support. Previously to find this answer one would have to go to the help site, search for information, and, then redirected onto another browse window.
Validation
When considering usability testing, I found the following metrics would exemplify the success of the chat bot design
Time on task
Time on task will demonstrate if the implementation of the chatbot is supporting clients enough to simplify task and reduce time on task
S U S
System Usability Scale will give quantitative and qualitative insights on how clients feel about the Chatbot’s usability
Heat mapping
Tracking use through heat mapping will directly show if clients are gravitating toward the Chatbot
<span data-metadata=""><span data-buffer="">Challenges &
Considerations
Challenges
Data
The lack of previous experience data made ideation a bit more difficult.
Primary Research
Having primary research results to synthesize could have made for a more intuitive experience
Time Constraints
Having more time would have allowed me conduct and gain primary research & testing insights
<span data-metadata=""><span data-buffer="">Accessibility
Considerations
-
Navigation
Consider keyboard navigation process to provide optimal navigation accessibility
-
Aria Labels
Consider aria labels to provide context and distinction to screen readers
Future Steps
-
Content Audit
A thorough content audit would allow for a starting point on how to govern and redistribute content through microcopy and other resources
-
Content Creation
Some content may need to be reimagined and some created completed from scratch to provide help and resources that are concise and easily digestible and accessible
Thank You
© 2021 All Rights Reserved.