01. project description
Conversational AI-based chatbot solution for the banking sector to help improve customer support, automate processes and provide better personal banking experiences to customers.
I worked in a cross-functional team as the sole designer to deliver a full end-end conversational-AI based chatbot application for banking.
02. project info
Role
App UI Design, User Testing, User Interviews
Timeline
September 2021 - October 2021
Team
Project Manager, Data Scientists, Software Engineers,
Product Designer (me), DevOps
03. the challenge
How Might We help banks provide customers with a more personalised banking experience and 24/7 support for customer queries?
Banking companies receive a large daily influx of inquiry call with a thousand more customers waiting in line to speak with a representative. There is no 24/7 customer support available to customers and if they were to employ human agents for this, would be a rather expensive cost for most banking companies. On the other hand, most banking application interface are either overly complex or outdated in their designs. Customers often times have to go through a maze to find how to perform tasks on their banking applications.
04. the goal
05. the solution
LugahBot is an AI-based banking chatbot integrated into banking applications serving as the user’s personal assistant, a financial advisor, customer support. BankBot uses natural language algorithm to understand what users enter as input while also using chat prompts to minimise typing for users. Features include
06. research
My research goal was to get an in-depth understanding of the current customer banking experience using their banking application chatbots, the shortcomings of these chatbots and their experience with banking companies customer supports. The information gathered from this was going to help form my design decisions and identify key requirements of customers in creating a better user experience for them. Some of the guiding questions for this research included:
01
How do people feel about chatbots?
02
What are some conversational AI practices?
03
What are some of the major tasks users perform on their banking applications?
04
What does a pleasant banking experience look like, for customers?
07. competitive analysis + the gap
I wanted to take a look at the existing chatbots offered by the banking companies in Nigeria to identify any gaps or shortcomings within these existing applications and areas of improvement I could target in creating a more meaningful experience for customers. I conducted a competitive analysis across 2 different bank chatbots. I identified the strengths and weaknesses of all 3 and identified the key services offered by all chatbots. The chatbots were tested on the following:
01
Personality and customization
02
Level of security and establishment of trust
03
Multi-lingual support
04
Response-time and reply accuracy
05
Features and functionality
06
Platform options
07.1. competitive analysis - findings
09. user interviews
I spoke to 3 users in the age range of 21 - 60. I wanted to understand their current frustrations and experiences with banking applications.
Transcripts from a few of the interviews:
Annette Black
I have to wait several minutes on the phone and endure being put on hold just to get an answer to a problem I have. It would be nice to able to talk to someone online. The whatsapp bot is supposed to help with this but last time, I didn’t get a reply till the afternoon of the next day.
Dianne Russell
The interface of my bank application is overly complicated. Even with the new re-design. I was hoping they would take out some things and present users with the most important features first. But they just keep adding more features and no way to hide them.
David Johnson
Opening an account is stressful for me. My daughter told me about some digital bank app I could use to save money conveniently. But had to go through so many form questions which got me tired as it required my 100% attention.
10. insights & opportunities
After speaking to some users and doing my online research, and analyzing the gathered information from the online research and user interviews, to help have a linear direction, I categorized the insights into these 2 categories.
97%
of task users want to carry out are centered around: fund transfers, account balance viewing, transaction history and rectification of card debit issues
60%
of customers are skeptical to release personal information to chatbots except the transaction is happening within the bank application interface.
82%
expressed an Expectation mismatch - chatbots are often times too stiff for customers thereby leaving a gap in not making the conversations fluid.
11. defining the problem
To better understand the users’ needs and align our efforts, I led two workshops with my team of five. The workshop focused on defining the problem space and getting a clearer picture of the project. This helped to keep everyone in the loop of things and define the problem(s) we trying to solve for.
Product Planning Done in FigJam12. problem statements
After the workshop, we had reached a solid direction for the product design with clear understanding of the specific problems to solve for. We defined problem statements and brainstormed on some directions by turning the problem statements into HMW questions.
13. the mvp: workshop #2
Working on this project, I was time-crunched because we were planning to have a ready product in a month. This was because we wanted to demo the prototype at the World GITEX Dubai.
To help prioritize features that would go into release for the demo, I led a second round of workshop with the team. The purpose of the workshop was to have team members vote on features we should prioritize and get buy-in from the CEO before commencing design.
The following features were prioritized afterwards:
14. design: user flows
Once we finalised on the core features we wanted to include in the demo, I came up with two possible user flows depending on the “why” a user might use the chatbot application. This was to help us map out the optimal path for users to get them started off with whatever their need might be.
15. the design
The users are mainly going to be using our chatbot mobile application for one major reason: to get answers to one of their problems
Given the short time for delivery, I centred my design around simplicity and functionality. I wanted to ensure the key tasks users perform with their banking applications are clear and easy to carry out in a conversational manner via the chatbot.
design decision(s)
For the design of user interface, I opted for a part conversational, part point and click. This is to give users some control, even though they will be talking to a robot, the experience didn’t need to totally feel that way. So I incorporated stop gaps for users to confirm or cancel certain processes. That way, users still feel in control.
From my research, I discovered that the younger generations preferred instant messaging over voice communication while the elderly may have trouble typing their requests. So I optimised my design to accommodate both preferences by providing an option for voice interface as well as a typing area.
Welcome Screen
Phone Verification of registered users to grant access to application
Main screen with access to transaction history and account information
15. the app scenes
Lugahbot onboards customers easily and allows easy upload of all documents required for customers to open an account with a bank.
Customers will always have questions when their money is involved. But most banking call centres and customer support lines are not available 24/7.
Lugahbot is able to provide customers with answers to any information they require or need clarification on. While also red0recting them to a knowledge base system for more information incase needed.
Funds transfer is the most performed task on mobile banking applications.
Lugahbot is able to help customers make fast and secure fund transfer to any beneficiary, both new and saved ones.
Funds transfer is the most performed task on mobile banking applications.
Lugahbot is able to help customers make fast and secure fund transfer to any beneficiary, both new and saved ones.
16. usability testing
At the GITEX Expo, I had the application tested by some users and a UX Researcher from a Fintech companies as a potential solution for onboarding older adults on their digital banking app. What worked for them include:
Users liked the conversation prompts because it made the conversation faster and didn’t have to go around looking for task buttons
Users liked that the chatbot was prompt in its replies without keeping them waiting
One UX researcher found the user onboarding experience smooth adn could see how conversational onboarding could be beneficial to the older generation who prefer voice interface to instant messaging.
17. result & impact
After the beta testing done at the GITEX Expo and having some Fintech companies and bank representatives use the app to get feedback and suggested improvements, we collated the results and proceeded to release the first version of the application in November as an enterprise solution with more updates to be made as we engage businesses on the platform.
5
of task users want to carry out are centered around: fund transfers, account balance viewing, transaction history and rectification of card debit issues
85%
of customers are skeptical to release personal information to chatbots except the transaction is happening within the bank application interface.
10. did i win something?
Despite being my first time working on conversation interfaces, I was happy for the opportunity to lead the product direction for such a product that we went on to demo at the Dubai GITEX Expo. Here are some of the learnings I took away from working on this product:
Using a MVP matrix to know which features to move ahead with and which to drop will go a long way in helping to make sure you have a linear direction in designing your product for initial launch. This will also help with meeting tight deadlines while still being able to deliver a quality product with good enough features to get the product out there.
Minimalism is key in making sure users are not overwhelmed by the interface design and worst still, confused as to what to do. Knowing the why and who of the product will ensure that we are designing for the right people and also solving their problem(s).
Research is key to launching any product: making sure there’s a market/demand for it, knowing similar solutions that exist and critically analysing the existing solution to make informed decisions in designing your product. This will not only save costs but also time and energy invested in the product.
Before getting a final product out there, several iterations happen. Iterations take a lot of time and can involve so many back and forth. One thing that will save a lot of time and these back and forth is getting feedback at every stage of the product design. Make sure all key players: stakeholders and team members are carried along through it all and adequate testing of solutions are done.
MORE PROJECTS
Revamped a mobile-first digital payment application for customers and 20 small POS business owners to enable cross-border payment across 5 countries.
Designed a foodprint value-label and web mobile application that empowers eco-conscious students to make sustainable food choices.
