Knowledge Sharing
To comply with the non-disclosure agreement, I have excluded and obfuscated any confidential information in this case study.
The
background story
The existing client is a multi-agency company who managed customer inquiries from multiple websites. They hired our chatbot to reduce the number of emails and calls they received every day.
Each agency will have its own chatbot and platform to handle agency-specific inquiries.
After a month or so, she found an interesting trend. Many of its customers' inquiries were actually common across the different chatbots. We realized there's a need for "common intent". But how do we then ensure such common intent's content is accurate and consistent across the different chatbots?
DISCOVERY
Knowing the situation from the BA and my product owner. I decided to conduct field studies and user interviews to find out:
Their existing workflow, for example: who can make the decision to add new intents?
Statistics, for example, what are some of the questions that were asked frequently?
How do they decide what intents to be shared?
What did they do to ensure that the answers are consistent?
To my surprise, even after adopting our chatbot platform, they are still using Excel sheets and email to manage their content and these are the reason why:
When the agency updates an intent, he/she has to notify the rest of the agencies to update their content.
Some intents are only valid for agencies A and B but not C & D
Our product is increasing their workload instead of reducing it.
PROBLEM
Our existing product wasn't able to support the clients' needs, we need to know:
How do we share intent across the different chatbots?
If the intents are being shared then how do we catered for agency-specific answers?
How do we handle conflicts between a private and shared intent?
CONCEPT IDEATION
Now, that we have a clearer idea of the situation. The product owner and I sat down to think of possible concepts that might resolve the issue.
Whiteboarding
I suggested doing some quick prototypes to test the concepts with the clients.
Prototypes
Feedback
SCENARIO MAPPING
With the feedback collected, I gathered the team of 8 developers, 1 QA, 1 BA, 1 Tech Lead, and 1 Product Owner to attend a scenario mapping session.
Scenario mapping helps the team to see the experiences that our users are having as they traverse through their activities. The relationship between task and experience provides a tremendous opportunity for insight.
Scenario Mapping
We then prioritize the items using a User Value vs Effort/Risk graph.
User Value vs Effort/Risk graph
USER FLOW & USER STORIES
Designing the user flow helps me consider the decision-making that the user needs to make at every point of the product.
User Flow
User Stories help the designers to fit our user personas into the context of the product we're designing and scoping out the tasks.
User Stories
PROTOTYPING
The following prototypes were created in a week for user testing
USER TESTING
The feedback from the user testing was analyzed for further iterations
User Testing Feedback
OUTCOME
The Buyer was very pleased with how we addressed the issue and decided to invite more agencies onboard the platform.
Eliminate the need to maintain a separate worksheet to manage intents sharing
The client no longer need to send emails to inform other agencies about intent updates