Centralized Help Center
Creating a central help center for FINRA’s various user groups
CLIENT:
FINRA
AREA:
AI Chatbot | Customer Support
YEAR:
2023
ROLE:
Lead Product Designer
problem
FINRA is an organization that safeguards the United States’ financial market and protect investors from fraud. FINRA oversees over 600,000 individual members and 3,000 member firms.
Financial markets or transactions can happen in an instance, leaving investors or investor firms concerned. It’s important that these users can easily find help at any time.
Originally, anyone looking to contact FINRA could do so by finding the Contact FINRA page. On that page, users have the option to contact FINRA via phone or email, schedule a call, or select a topic area to be redirected for further support.
Users are unclear of where to go for support and often face extended wait times when trying to connect with live agents.
my role
• Create comparative analysis of existing support chatbots
• Devise user scenarios for user interactions with chatbots
• Create high-fidelity design of chatbot experience
• Share and iterate on high-fidelity design with stakeholders (including business team, engineers, and product managers)
• Document design specs for the final design
challenges
1. A Help Center that can be launched autonomously on any page in FINRA's enterprise platforms. When I was tasked with designing an AI support chatbot experience, one of requirements was that it had to be easily accessible on the FinPro platform, Gateway platform, and the .org website, accessed by hundreds of thousands of users. After experimenting with the placement and component type, I created a Help Center in the top navigation bar, right-aligned and next to the Recently Viewed and Notification alerts. This Support Center provides multiple options for the user to receive help including: chat with an AI agent, speak with a live agent, schedule a call with a live agent, and submit a ticket.
2. A feature that allows users to schedule a call with a live agent directly. From survey feedback, we found that the majority of users are hesitant to speak with an AI chatbot and would prefer the assistance of a live agent. So, for users who feel they must speak with an agent, we provide an option for users to go this route. However, as the AI chatbot continues to exist and people use it, the AI chatbot will continue to improve with machine learning.
solution
An omni-channel support experience where users have a variety of avenues to find help including:
• AI chatbots
• Live agents
• Guided resources
• Feedback submission forms
results
48% of 20,000 inquiries were resolved by the AI chatbot.
48.5% of ratings for the AI chatbot were positive.
Significant reduction in the frequency of calls and chats with live customer support agents.
learnings
1. Skepticism towards AI agents persists. Users are hesitant to engage with AI chatbots due to past frustrations with inaccurate responses or unsatisfactory interactions.
2. Desire for live agent support. Despite improvements in AI chatbots, users expressed a preference to connect directly with live agents, whether through chat or phone, for more personalized assistance.
