To adhere to the confidentiality agreement, I have removed and disguised all sensitive information in this case study.
Background Story
Our client, a conglomerate operating several agencies, sought to diminish the daily influx of emails and calls through the deployment of our chatbot services. Each agency was equipped with its tailored chatbot and platform to address unique inquiries.
However, within a month, an intriguing pattern emerged: numerous customer inquiries were not unique to one agency but common across the spectrum of chatbots. This revelation highlighted the necessity for a unified approach to these "common intents." The challenge then became ensuring the accuracy and consistency of responses to these shared inquiries across all chatbots.
Discovery
Upon recognising the challenges faced by our client, I initiated a series of field studies and user interviews to delve deeper into:
Their current operational processes, specifically identifying the decision-makers responsible for introducing new intents.
Key statistics, including the most frequently asked questions by customers.
The criteria for determining which intents should be shared across different agencies.
Strategies implemented to ensure consistency in responses provided by the chatbots.
Unexpectedly, I discovered that despite adopting our chatbot platform, the agencies continued to rely on traditional methods such as Excel sheets and emails for content management due to the following reasons:
Whenever an agency made updates to an intent, it necessitated a communication to other agencies to synchronize their content accordingly.
Certain intents were applicable only to specific agencies (e.g., Agencies A and B) but not to others (e.g., Agencies C and D).
Instead of alleviating their workload, our product inadvertently contributed to its increase.
Problem
Our current product couldn't meet the requirements of our clients. We need to understand:
How can we distribute intents among various chatbots?
If intents are shared, how do we provide answers specific to each agency?
What approach should we take to resolve conflicts between private and shared intents?
Concept Ideation
Now that we have a better understanding of the issue at hand, the product owner and I have collaborated to brainstorm potential solutions.
Whiteboarding
I proposed creating some rapid prototypes to trial the concepts with our clients.
Prototypes
Feedback
Scenario Mapping
After collecting feedback, I assembled a team comprising 8 developers, 1 QA analyst, 1 business analyst, 1 technical lead, and 1 product owner for a scenario mapping session.
Scenario mapping enables the team to understand the experiences of our users as they navigate through their tasks. This link between task and experience offers significant potential for insights.
Scenario Mapping
We then prioritised the items using a User Value versus Effort/Risk graph.
User Value vs Effort/Risk graph
User Flows & User Stories
Creating the user flow allows me to take into account the decisions that users must make at each stage of interacting with the product.
User Flow
User stories assist designers in integrating our user personas within the context of the product we are designing, helping to scope out the tasks involved.
User Stories
Prototyping
The prototypes were developed within a week for the purpose of user testing.
User Testing
The feedback gathered from user testing was examined for subsequent iterations.
User Testing Feedback
Outcome
The Buyer was highly satisfied with our resolution of the issue and chose to invite additional agencies to join the platform.
Removed the necessity for maintaining a separate worksheet for managing intent sharing.
Clients no longer have to send emails to notify other agencies about updates to intents.
Комментарии