ROLE
Product Designer
LAUNCHED
June 2024
SCOPE
iOS and Android
OVERVIEW
The expanding user base and the strain on support staff due to non-automated tasks are leading to burnout and decreased efficiency in troubleshooting simple issues. Implementing an AI chatbot to automate these tasks would enhance user satisfaction, leading to better reviews and higher rankings in app stores.
70% of tasks
Handled by the A.I powered chatbot
40%
increase in positive user reviews in app store
Saved $25,000
due to basic tasks being handled by the chat bot
What was wrong?
Due to slow response times and insufficient self-help resources, users experienced dissatisfaction when seeking assistance, highlighting the need for improved support solutions. With the expansion of the user base and the increased capabilities of artificial intelligence, it's crucial for the business to automate support to enhance scalability and efficiency.
What were our goals?
COMPANY GOAL
Increase organic ranking of app in store's via increasing positive reviews based on customer support. This would allow the app to be suggested more and rank higher in searching.
PROJECT GOALS
Provide instant, accurate responses to common queries.
Assist users in navigating the platform and executing trades.
Offer personalised recommendations based on user behaviour and preferences.
Seamlessly escalate complex issues to human agents when necessary.
Ask for reviews after questions and issues have been resolved.
Reduce time support team have to answer mundane tasks.
Analysis and workshops
Competitive analysis
This helped me get a clear image of the landscape in the Australia market. I was able to make clear notes on what worked and what did not work.
Workshops
After doing the analysis it gave me a solid foundation of questions and exercises we could do to get through the the main issues that needed to be addressed when designing the chat bot.
Information architecture and flows
Users had to wait on law firm to approve
There were a number of problems with the investigative search flow. As it could take two weeks and the possible firm may not have any data on the will I needed
Billing had to be split in to two different stages
First a non refundable deposit that had to communicated clearly.
Second is final payment to view results when they were in.
Conversation design
Dialogue Scripting
The next area of focus was dialogue scripting, which involved two key steps:
- Writing scripts for different scenarios
- Ensuring the language was natural and engaging.
This approach helped customers feel at ease and confident that their questions were being answered.
Handling Edge Cases
The next focus was on addressing edge cases, as identifying specific problems can be challenging. I planned responses for unexpected user inputs, created fallback messages, and implemented robust error handling.
The key was to ask for more direct information to ensure questions could be accurately answered.
All aboard the feedback train…
Upon testing with support staff there were some key ideas and points they pointed out.
Needed popular answers based on their account and settings
Needed a like and dislike button on responses so answers could be improved.
Conclusion
The AI-powered chat bot project for Tiger Brokers successfully addressed the company's challenges by providing a scalable, efficient, and user-friendly solution for customer support. The project's success underscores the importance of integrating advanced technologies and user-centered design principles to enhance service quality and drive business growth. Tiger Brokers continues to refine and expand the chat bot's capabilities, leveraging ongoing user feedback and technological advancements to further elevate the customer experience.