You're running late for work, and you need to transfer some money to your rent account. You don't have time to log into your bank's website or call customer service. But you do have time to say, "Hey, AIBA, transfer $500 to my rent account."
AIBA's conversational AI banking app understands your request and transfers the money instantly. You're on your way to work, and you're already feeling more relaxed.
This is just one example of how natural language models can be adopted in fintech applications. By enabling customers to interact with their banks in a natural, conversational way, the experience can make banking more convenient, efficient, and personalized.
With conversational AI, customers can access banking services 24/7, get personalized advice, and be confident that they are getting the correct information. Banks can save money by using conversational AI to handle a large volume of customer inquiries, freeing up human agents to focus on more complex tasks.
Plus additional statistics that illustrate the benefits of conversational AI for consumers:
I joined the team at AIBA as a sole UX designer for this very reason. My goal was to make the banking experience more intimate for customers, and easy for everyone to use.
5 Weeks
4 (Full stack developer, AI architect, UX Designer, UI Developer.
Summer 2023
Project Goals:
Project Objectives:
Project Requirements:
Project Deliverables:
Project Risks:
Project Mitigation Strategies:
"AIBA has to feel like it's introducing a new concept into fintech. Not reinventing the wheel, but directing the experience to a new path where having conversation are more important than navigating through user flows."
In 2022, the global conversational AI market was valued at $10.7 billion, and it's expected to reach $29.8 billion by 2028. The banking and financial services sector is one of the largest adopters of conversational AI, accounting for 25% of the global market in 2022.
Conversational AI is becoming increasingly popular in banking apps. Here are some statistics that illustrate the growing adoption of conversational AI in banking:
These numbers are impressive and further validates the growing need to adopt conversational AI in banking apps.
I conducted a simple survey with some potential users to help me break down some of the areas in mobile finance that they experience as well as solutions they're hoping to see. I kept my interview small with only 4 interviewees. For the sake of their privacy, we kept their identities hidden.
The surveys helped me realize the sentiment of my interviewees and how they feel overall about conversational AI in banking. Among all the responses, I noticed that a majority of the interviewees have an optimistic attitude about it.
The vast majority of customers aren't fond of chatbot experiences being in their fintech apps, typically. The use of conversational AI is new to the industry and customers often ask questions relating to their user data. Questions like:
"Is my information safe?"
"Am I talking to a live rep or an AI?"
"Does the AI remember my info?"
"Does my info get stored somewhere safe? How am I sure?"
Phase 2: Discover & process info
By identifying my users pain points early on in the process, I was able to list out the initial features to add into the product/service. This method helped me build for the users needs continuously.
Based on my findings, customers face:
Where there comes a problem, it requires a solution. I followed up on the customers pain points and drew up some user-need statements in order to get a deeper sense to our customers needs.
I created user personas to better illustrate my customers. For me, user personas are like characters in a story. They represent the different types of people who will use the product or service. Now, I'm able to get to know my users better and understand their needs and wants. This information can then be used to design a product or service that is user-friendly and meets the needs of its users.
I narrowed down the info collected into bullet points to better characterize customers use cases:
Some of the reasons they'll enjoy conversational AI banking app is because:
By understanding the challenges and goals of each persona, I was able to design an app that's tailored to their needs. For example:
"John Smith is a busy professional who values convenience and efficiency, so the app should be designed to be easy to use and navigate. Jane Doe is a retired woman who is not very tech-savvy,"
so the app should be designed to be easy to use and have clear instructions.
I'm going to prioritize features that are important to users. Using the needs assessment that I conducted, I'm able to make sure that the app includes the features that are most important to them. For example, John Smith would appreciate the ability to ask the app questions about his finances and get personalized advice. Jane Doe would appreciate the ability to manage her finances without having to go to the bank.
Furthering my connection with end users, I was able to connect with their problems more intimately. Come see how I dove into my design strategy.
A chatbot banking app can start by welcoming the customer and asking how they can help. This will make the customer feel welcome and appreciated, and it will also help the app to quickly identify the customer's needs.
For example, the app could say something like:
"Welcome to the bank! How can I help you today?"
It should also be able to answer customer questions in any language. This will help the customer to get the information they need quickly and easily, without having to wait on hold or speak to a live representative.
For example, the app could answer questions like:
"What is my account balance?"
"When is my next bill due?"
"How can I transfer money to another account?"
The conversational AI banking app should be able to help customers with their transactions. This could include things like:
* Initiating a wire transfer
* Paying a bill
* Depositing a check
* Transferring money between accounts
It should be able to provide personalized recommendations to customers. This could include things like:
* Suggesting new products or services
* Offering discounts or promotions
* Helping customers to save money
The conversational AI banking app should be available 24/7. This will allow customers to get help with their banking needs no matter what time of day it is.
I was tasked with creating wireframes, a low-fidelity sketch of the user interface. These wireframes helped me to understand how a typical conversation between a customer and the chatbot would flow. I also used the wireframes to experiment with different tones, responses, and suggestions that the chatbot could use to help customers complete tasks.
Once I had a good understanding of how the chatbot would work, I began user testing. I recruited a handful of potential users and asked them to interact with the chatbot. I observed their interactions and collected feedback. This feedback helped me to refine the chatbot's design and make it more user-friendly.
After several rounds of user testing, I was satisfied with the chatbot's design. I moved forward with the visual design.
Information Architecture ( NDA)
User Flow & Guidelines ( NDA)
Brand Design Guidelines & Pattern library (NDA)
Design specifications & User Stories ( NDA)
While I was designing AIBA; I knew that the home page was going to be crucial. After all, that's the first thing that customers would see when they opened the app. I didn't want it to be overwhelming, so I decided to simplify it.
I started by listing down the most important things that customers care about when using a fintech app. These included things like account balances, recent transactions, and upcoming bills. I then made sure that the most important items were prominently displayed on the home page.
I also wanted to make the home page more personable, so I added a welcoming message from AIBA. This was a small touch, but it made a big difference.
Finally, I didn't want to eliminate the user interface entirely. After all, customers still needed to be able to access their financial information. So I added a "Slide left" feature. When customers slide left from the home page, they were directed to a dashboard that showed a visual view of their financial activities.
I was really happy with how the home page turned out. It was simple, yet it was also informative and personalized.
I knew that one of the most important things for customers was being able to access their bank account information quickly and easily. That's why I created the "Info button." This button was always available in all parts of the app, and it gave users access to a list of commands of common prompts they could use to access their bank account information. For example, if a user wanted to see their balance, they could simply click on the "Info button" and select the "Balance" command.
I also thought it would be helpful to keep the "Info button" by the text box. This way, users could always pull up the "Commands" tab if they didn't know how to phrase their request.
I also had the choice to make it a text-only interface; However, I knew that this would make it difficult for customers to visualize certain banking activities, such as paying bills or viewing their latest transactions.
To address this, I decided to add “card designs” to certain commands. This allowed customers to see a visual representation of their data, which made it easier for them to understand and interact with it.
For example, when a customer asked to pay a bill, the chatbot would display a card with the bill’s details, such as the amount due, the due date, and the recipient. This made it easy for the customer to confirm the payment and complete the transaction.
I ensured that the vast majority of the experience was text-based, as this is what most customers are used to. However, by adding “card designs” to certain commands, I was able to give customers a more dynamic and engaging experience.
When designing the onboarding process for a banking app, I had to make sure that it was both secure and user-friendly. I knew that security was a top priority for customers, so I decided to stick with a traditional process.
Users are able to login via Gmail or provide their own email address. Once they have created an account, they are required to undergo a two-factor authentication process(2FA).
I also added a progress bar to the top right of the screen. I believe that these measures will help improve the conversion user logins and reduce churn rates.
AIBA uses Plaid as the payment integration API. When designing the "Add Bank Accounts" screen, I realized that the search bar would be difficult for customers to tap when using an iPhone 11, 12, 13, or 14. The search bar was originally located at the top of the screen, but this made it difficult for users to reach with their thumbs.
To address this, I decided to move the search bar to the bottom of the screen. This made it much easier for users to tap, and it also improved the overall usability of the screen.
I thought it was important to give users the option to select which parts of their bank accounts they would like to pull data from. After all, not all users need access to all of their account information.
To address this, I created a process that allows users to select the specific accounts that they want to connect to the app. This process is located in between the "Logging into bank" and "Verifying the completion of the log-in" steps.
I made sure to offer suggestions to customers in between conversations with AIBA. I thought this was important because it helps the chatbot learn about the customer's habits and preferences. This, in turn, allows the chatbot to offer more accurate and relevant suggestions in the future.
For example, if a customer repeatedly asks about their balance, AIBA might suggest that they set up a balance alert. Or, if a customer frequently pays their bills late, the chatbot might suggest that they set up bill reminders.
By offering suggestions, the chatbot can help customers save time and money. It can also help customers stay on top of their finances and make better financial decisions.
I believe that this is a valuable way to train AIBA. It helps the chatbot learn about the customer and provide them with the best possible service.
I designed AIBA, that it may typically ask customers to confirm their phone number using two-factor authentication (2FA) for large wire transfers or money transfers between accounts. This helps to prevent fraudulent banking activities and reassure customers that it is safe to conduct such transactions with a chatbot.
When you click on "Continue to Netflix," you will be taken to a third-party website. The privacy practices of this third-party website may differ from AIBA's privacy practices. We recommend that you review the privacy statement of this website before providing any personal information. AIBA is not responsible for the privacy or security of this third-party website.
I also laid out the steps as to what needs to occur in order to successfully pay a bill electronically using a banking chatbot.
It's a common approach for customers to call their representatives when they lose their debit card; Therefore, I added it to the list of commands that AIBA offers in order to turn the Fraud reporting feature into a conversational UI interface. My goal is to allow the chatbot to make the process as easy as talking to a live representative, but via text. I designed this by emphasizing the main questions that live reps ask when reporting a card lost or stolen.
I understand the security measures it takes to ensure the customer's trust is kept. Therefore, I limited the questions to 4 input questions, that gives AIBA certainty that the customer is the person behind the lost card. I've also implemented an ETA at the end of the experience for customers to track their card delivery. I find this design very useful as the first set of prompts because it allows AIBA to understand how to deal with these cases, and remain diligent to find unique ways to help customers without compromising their account. This gives us valid data to configure ways how to improve this experience later on in the journey.
I tested AIBA with a few set of customers during the testflight phase, and received positive and negative feedback. 60% of users or 3/5 users said that the app should get smarter over time, and respond more emotionally to encourage positive reinforcement during times when customers are facing hardship with making financial decisions. 80% of users enjoyed sending/receiving and transferring funds but will like to have the option to send a message to the person they're sending funds to. All users say that they can see conversational AI being an integral part of their lives, and an easy way to make quick financial choices on the fly.
