SCB Connect: AI Advisory Chatbot
Improving usability and boosting engagement for a fund advisory chatbot at Siam Commercial Bank (SCB).
My Role
As an experience design intern, I was fortunate enough to have led 2 projects I initiated which includes new in-app broadcast features to improve user engagement and the enhancement of the fund advisory chatbot conversational flow and UI. Throughout my internship, I utilized journey mapping, competitive analysis, survey design, user interview, facilitated in-person workshops, high-fidelity prototyping, and SUS testing.
Project Scope
I worked on this project from June - July 2024 as a part of an internship with Siam Commercial Bank in SCB Connect, the digital banking team.
Project Outcomes
+ Revamp of the conversational flow and UI of 6 chatbot intents improved the SUS score of the AI Advisory Chatbot by 15%, which raised the usability from poor to good.
+ Received positive feedback across SCB Connect team.
+ Improved UI and navigation flow was shipped in November 2024 version update.
Background
AI Advisory Chatbot is a new feature on SCB Connect which allows users to naturally interact with fundamental fund information by querying their desired funds.
The overarching goal is to improve the usability of the AI Advisory Chatbot to allow for more natural conversations between users and the chatbot as well as to modernize the UI, making components more visually appealing.
Challenge
After conducting in-depth competitor research about investment chatbots among other leading Thai banks, I took a deeper dive into SCB’s AI Advisory Chatbot to identify gaps and understand what could make SCB stand out.
I gained domain-knowledge by familiarizing myself with the chatbot through generating my own prompts and using intents provided by the chatbot developers. This was an important step to put myself in the user’s shoes and test the chatbot’s limits. After thorough testing, I quickly realized many aspects, both content and interface-wise, of the chatbot responses that were confusing and frustrating:
Product Overview
Too many colors make it hard to focus on specific information
Center alignment of text is difficult to scan
Card is too long (doesn't fit in one screen)
Card is too long - users have to scroll a lot
Dense text makes it hard to scan for key information
Purple buttons don’t look clickable
Order of information doesn’t make sense
Doesn’t supply sufficient details instantly
Product Recommendation
Ideation
After identifying content and layout issues, I narrowed my focus to six main chatbot intents for improvement. I collaborated with chatbot developers to determine which intents had the highest interaction rates and aligned with product managers to prioritize those most crucial for encouraging customer investment through SCB. Throughout, I kept in mind that the main goal of the AI Advisory Chatbot was to provide added value to SCB customers, ensuring they continue using SCB Connect.
The 6 chatbot intents I focused on are:
1 - Entry Point
"What is AI Advisory Chatbot?"
2 - Fund Overview
"SCBSET50"
3 - Fund Recommendation
"Recommend me some tech funds"
4 - Funds for Beginners
"I want to invest but don't know where to start"
5 - Fund Information
"How much percent does PTT have in SCBSET50?"
6 - Current Market Situation
"How is the stock market in the US?"
Iteration
Before jumping into the redesign, I spent a lot of time understanding the current use cases, mapping out potential use cases, and looking at the chatbot from a broader entity as the chatbot as a product connected to a lot of other products from the bank like the SCB EASY App and SCB's website. Many iterations were made to the UI design after design feedback from another UX designer on the team and technical feedback from chatbot developers.
Design system
Since there was no previous design work done on the UI of the chatbot, another UX designer and I built the foundations of a new design system from scratch that aligns with SCB’s branding in order to create a more modern and consistent look for the chatbot.
System Usability Testing (SUS)
Will these changes make querying for fund information more effective?
As I reached my final iteration, I completed System Usability Scale (SUS) testing with the digital banking team in order to gain an objective understanding if the changes I made to the navigation flow and design actually improved the user experience when interacting with the AI Advisory chatbot. After the test, the SUS score raised by 15%, demonstrating the improved effectiveness of the new UI and conversational flow.
Proposed Solution
By the end of the internship, I was proposed 6 revamped navigation flows and UI to the 6 chatbot intents.
Disclaimer: The screens presented in the mock-ups below are not final. The designs are aware of Thai context as the primary target audience is the Thai population.
Entry Point
Get to know AI Advisory Chatbot’s features through 2 entry points:
1 - A short trivia game about investment vocab
2 - An easy-to-understand summary of investment vocabulary
Overview
Type up to 5 fund names to get information about it’s risk score, NAV value, trailing return, and more!
Product Recommendation
Want to invest in industry-specific funds but don’t know where to start? Just type something like “Good tech funds” and you’ll get a list of 5 recommended tech funds by SCB.
Product Information
If you want to ask in-depth questions about a specific mutual fund, AI Advisory Chatbot will provide you with the answer, as well as allow you to give you the option to invest via SCB EASY APP (purple button) or look at the fund fact sheet (light purple button).
Current Market Situation
Learn about how the stock market is doing in different countires, including the US, China, Europe, India, Thailand, Indonesia, South Korea, Vietnam, and Japan!
Version Update as of November 2024
What I learned
Working with chatbots in banking taught me the critical role of internal compliance, including when to display disclaimers, even if they seem redundant or misaligned with design. Additionally, data demonstrated that hard-selling bank products is ineffective, so when collaborating with teams unfamiliar with UX, I emphasize how a strong user experience not only boosts engagement but also drives business objectives by subtly persuading customers to buy through thoughtful design.