Back to Blog
According to the dictionary, a chatbot is a computer program designed to simulate conversation with human users, especially over the internet. The conversation is conducted via text or text to speech software, using messaging platforms as well as proprietary apps. For the last 10 years, machine learning and artificial intelligence have made it possible for chatbots to hold intelligent real-time one-to-one conversations at scale. Right now, the chatbot market is growing at a staggering 34.75% CAGR, and is expected to reach $100 billion by 2026.
“Chatbots everywhere” is not just a buzz-phrase. It’s the new normal for 2021 -- and beyond.
The banking sector traditionally has been a leading adopter of cutting-edge technologies because of its needs for speed, trust, and communication. Banks have been using chatbots since the early 2000s, primarily to build customer relations by cognitively learning about customers’ thinking and providing instant responses. Let’s look at the latest chatbot trends in the banking sector and see how some of the more digitally mature banks use chatbots today.
Chatbots are easier to use than a traditional banking app or IVR (interactive voice response system). Customers can access the chatbot via messaging apps they’re already using, such as Facebook Messenger or WhatsApp. The quality of chatbot interactions rapidly improves as the AI learns more about each customer’s preferences and needs. According to figures from Salesforce, 69% of customers prefer to interact with brands via chatbots.
Image Source: Net Finance
Example: Bank of America’s chatbot named ‘Erica’ provides customers balance data and credit report updates and helps them pay bills and perform simple transactions.
Chatbots require no salary, lunch breaks, or sick leave. They can be online 24/7 on multiple platforms and are inexpensive to develop and train compared to their human equivalent. Cloud based chatbots can be customized with a pay-as-you-go price. They eliminate the need of having local data storage or servers, further bringing down the TCO and operating costs.
Image Source: The Week
Example: HDFC’s chatbot EVA (Electronic Virtual Assistant) saves cost by providing branch addresses, IFSC codes, loan information, and other reports without requiring human employees.
A banking chatbot can help open new accounts by providing website visitors with the information they need and pointing them in the right direction. It can reduce the workload on human employees and deliver more customer satisfaction by reducing the response times.
Customers can add Amex chatbot to Alexa as a “Skill” (Source)
Example: Banking brands have reported an up to 600% increase in the number of leads by integrating chatbots on their websites.
Chatbots can collect better-quality data more easily because customers willingly provide information when they are engaged in a conversation. Banks can use this data to personalize their marketing campaigns and refine their services, increasing customer loyalty and satisfaction. Customer data is safer with chatbots compared with other channels.
Capital One’s Eno SMS chatbot manages customer credit cards and bank accounts. (Source)
Example: Bank of America, American Express, MasterCard, and Capital One have integrated chatbots into their service infrastructure. These banks and financial institutions collect and use relevant customer information to design more targeted products and services.
Advanced AI based banking chatbots can access and interpret all of customers’ data including their spending habits, credit scores, and more. They can use this data to provide financial advice. They can set and manage budgets, tell customers where they’re spending their money, and give advice and recommendations for better financial management.
Example: HSBC uses a ’female’ chatbot Amy; it continually learns and enriches her knowledge to keep up with the increasing number of broad queries. (Source)
A banking chatbot is an AI based software application that can converse with a real person to answer a question, solve a problem, or provide a recommendation. Automating customer conversations allows banks to reduce response times and handle a larger number of customers without hiring more employees.
A chatbot may be able to hold conversations in three ways.
Chatbot programming is similar to autoresponder programming. A simple chatbot can be designed by anticipating the questions customers are going to ask and pre-configuring the answers into the app. The answers can be triggered by the software based on keywords that a customer query contains.
Artificial intelligence including natural language processing (NLP), natural language generation (NLG), and natural language understanding (NLU) have revolutionized today’s chatbots. These technologies allow banking chatbots to understand and generate language structure and also infer intent from language even when the written or spoken language has an incorrect structure. An AI chatbot can understand questions the same way a human does. However, the AI chatbot needs to be “trained” before it can understand customer questions correctly and provide the right answers.
The pre-launch training starts with huge datasets and continues with live customer conversations. Banking chatbots require additional training in learning bank-specific tasks. For example, they need to learn that a branch doesn’t mean a tree part but a building. Unlike most people think, the actual coding is a very small part of AI. Most of it is learning through data and experiment.
Image Source: The Financial Brand
If you work at a financial institution or a bank and have been tasked to integrate a chatbot into the bank’s operations, customer services, or other internal/external banking processes, there are basically two ways to go about it.
Which route you take will depend upon the nature of the job, your time frame and budget, and whether pre-trained chatbots can fulfill the requirements you have. The problem in going with a general purpose AI platform is that the entire burden of designing and training the chatbot rests on the you and your bank. The investment and scope of such a project expands continuously, as the AI requires constant machine learning. This may be the right approach if your bank is looking to be a pioneer by making a groundbreaking innovation—such as Erica by Bank of America.
The alternative is using a pre-trained banking chatbot application managed by an external provider. A managed chatbot is a more cost effective option by far. It allows the bank to place all the training, infrastructure, machine learning, and integration tasks to the provider. Managed chatbots use cloud computing and are available at a pay-as-you-go price, which makes them the best option for small and medium banking institutions. Let’s take a closer look at some of the best managed chatbots in banking.
KAI is a US-based conversational AI platform created to master the language of banking and finance. It is a chatbot and virtual assistant trained and ready to meet the demands of modern banking customers. KAI can be deployed on a bank’s messaging, mobile, and web platforms. The chatbot, with its built-in banking-specific knowledge, can assist customers with payments, transactions, account insights, and personal financial management. Several major banking institutions use KAI, including DBS, Standard Chartered, Mastercard, and RBC.
The London-based Personetics Assist chatbot is a self-service AI application designed specifically for the banking industry. The chatbot can use NLP and predictive analytics to comprehend and anticipate customer questions and offer appropriate information and recommendations. Assist integrates with a bank’s mobile app, as well as messaging platforms such as Facebook Messenger, and Amazon Alexa. The chatbot can also send payments, set up appointments, change passwords, and perform other common banking tasks, as and when requested by a customer.
Headquartered in British Columbia, Vancouver, Canada, Finn AI focuses on AI for the banking industry. It is a conversational tool that uses NLP to bring personalized experiences to a bank’s customers. The built-in machine learning processes allow the app to recognize banking related customer queries. Finn AI is being used by Bank of Montreal, which has rebranded the chatbot as BMO Bolt. The Bolt is configured to answer 250 common questions relating to the Bank’s products. ATB Financial, another Finn AI user, claims to have made the chatbot available to more than 1 million of its personal banking customers.
Michigan-based AI startup Clinc offers several banking applications. These include Finnie Personal, a retail banking product, and Finnie Wealth, a wealth management chatbot. The company claims banking institutions can deploy both the apps on multiple platforms and channels, including IVR, mobile, messaging apps, web, and chatbots. Clinc’s AI banking products can process voice and messaging data to understand customer queries and problems, and provide the appropriate response. Banking customers have a personalized experience while performing transactions and accessing the information they need.
Kore.ai is an enterprise-focused AI solution provider headquartered in Florida. The company offers chatbot and virtual assistant applications for employees and customers. Kore.ai’s employee experience products include IT Assist, HR Assist, and WorkAssist—software applications that assist employees resolve their issues and queries in the areas of IT, HR, and day-to-day working. Customer experience products include contact center automation, digital customer service, and Bank Assist—an intelligent VA that responds to regular retail banking queries through personalized conversations.
When the Coronavirus pandemic hit, customers looking to defer mortgage or credit card payments, collect unemployment, cancel airline flights or locate missing shopping orders had to endure record wait times to speak to customer service agents. The Chase banking website and app warned of “Extremely long wait times if you call us.” Citibank’s customers had to wait for as long as three or four hours, at times. Jolted by the crisis, banking institutions turned to chatbots en masse. The phenomenon led a spike in the development of AI, NLP, and Machine Learning tools.
Today’s intelligent chatbots and VAs use human-like conversations to understand, anticipate and respond to user requirements. They grow smarter with each interaction with humans. It seems the Singularity—a hypothetical point in time at which technological growth becomes uncontrollable and irreversible—may not be too far away. As far as you and your bank are concerned, the time to act and start integrating chatbots into your business is now.
The art is building the right growth campaign. The science is in the results.