ANALYSIS. STRATEGY. RESEARCH. DESIGN. INTERFACE DESIGN. ANALYTICS.
AI Chatbot
Lloyds virtual assistant helps make everyday banking easier
Lloyds Virtual Assistant integrated human and machine learning to create an improved model of customer communication. One which provides 24/7 support, is accessible on different devices, easy to use, more inclusive, and adaptable to different customer physical and information needs.
Design features include automatic recognition of emotions, bereavement, and anger, which when detected diverts the call to dedicated customer support teams. This easy-to-use communication channel is valued by customers and invaluable for those who find it difficult to travel to a branch.
Project goal
The goal was to provide better service, not replacing people in our call centers. The benefit of a bot is that it’s available 24/7, so the customer doesn’t have to wait. Another benefit is that the bot can be a very efficient first point of point. If the bot cannot solve or answer a problem, it will refer the customer to a relevant support agent. The agents can focus more on advice, less on answering repeat queries. Sounds a bit mechanical and boring? Bear with me…
My Role
I worked as the sole designer on this project as a UX designer and Customer journey expert. I set out to explore the optimal banking conversational experience.
This experience needed to integrate within existing systems, processes, and platforms on the back end, while delivering a heightened, delightful user experience on the front-end that would drive user engagement to deliver on several mission-critical business metrics and KPIs.
-
Design a complete set of conversational UI assets
-
Create a conversation script
-
Handle most common types of conversational randomness (meaning f***s and dfdffjhfdjhfkfs)
-
Convey the brand’s character (also by using company’s brand hero, Cody)
Accomplishments
Reduced Call Volume -
Increased Customer Satisfaction
Freeing up human agents at contact centers to address complex inquiries allows representatives to take their time and provide a better service.
Ease of use.
Intuitive and easier to use Chatbot. There is no download required and the experience can be personalized over time through machine learning.
Conversational
Interface studies have already shown that people find phone calls tedious and slow compared to instant chat systems. Customers are looking for speed and personalization when interacting with their financial institution. At the same time, they don’t want the impersonal experience of emails or online forms. Customer support chatbots offer a mix of both live conversation and speed.
Financial advice
The chatbot has access to all of a customer’s data. It can track spending habits, provide credit scores, set and manage budgets, and tell the consumer where they are spending their money. This allows for AI-based recommendations and advice for better money management.
New account generation
If a prospect visits a bank’s website and is greeted with a polite ‘hello’, invited to ask questions about a product, and pointed in the right direction to where they can find relevant information, the engagement is much more likely to result in them opening a new account.
24/7 digital support
A 24/7 instant chat feature is something that is expected by the increasingly digital ‘always-on’ consumer of today. This means that banks with chatbots built into their websites have a competitive advantage over others and are more likely to attract and retain customers.
Example flows:
Opening
The first message your chatbot sends to the user is more important than you probably realize. Like our daily interactions with other humans, we’re bound to develop strong initial impressions when interacting with a chatbot for the first time.
The chatbot greeting provides an opportunity to create a strong first impression that will hopefully bring in more recurring users.
Effective Chatbot Buttons
Buttons are often sent during a conversation by the chatbot for a person to click on. Such buttons allow users to view recent transactions for specific accounts just by clicking on the button - the user can select their query or issue. Similarly, a button can be offered as a real-time access point to another screen that the visitor was trying to go to. This means that buttons can actually do a lot of positive things for customer interaction.
Chatbot failing to understand a user's messages
When it comes to this, the bot can try to understand the user’s intent by asking them to rephrase something. However, if it’s too frequent or repetitive, people get frustrated and the bot can do more harm than good.
Still, it is challenging to train the chatbot to be ready for every possibility. Therefore, a much more reliable option is to train the bot to recognize its limits. That way, instead of failing to understand and help the user, it will be able to assess when a specific user request needs bot-human handoff. In that instance, the bot can simply say:
“I’m not able to solve this issue on my own. Please wait while I transfer you to one of our human agents who is sure to be of help!”
Chatbot Analytics: Essential Metrics & KPIs to Measure Bot Success
In-conversation feedback
Satisfaction score - The customer satisfaction KPI evaluates the user satisfaction level with bot conversation. At the end of the conversation, the chatbot can ask users a simple yes/no question about their experience. Also, I have introduced various ways for users to express satisfaction or dissatisfaction with the bot, including a scoring system (10-5 score).
-
“Was the chatbot’s response appropriate to the user’s query?”
-
“Did the user ultimately accomplish at least one goal of the conversation?”