Banking Chatbots – Comparing 5 Current Applications

Ayn de Jesus

Ayn serves as AI Analyst at Emerj - covering artificial intelligence use-cases and trends across industries. She previously held various roles at Accenture.

Banking Chatbots - Comparing 5 Current Applications

Developments in artificial intelligence have had global banks recently integrate online chatbots into their websites and mobile apps. Smaller, newer banks are following this example. The gaining popularity of chatbots could be considered surprising for an industry that handles other people’s wealth and perceives security as top priority.

However, today’s tech-savvy banking customers have begun to embrace the technology so that banks see this trend as an opportunity to provide a convenient 24-hour customer service without added labor costs, with the potential to open new revenue streams.

Traditional banks with an online presence usually require customers to seek out information through a phone call or a search in the bank’s website.

AI is further empowering banking institutions and retail banking customers alike by taking massive amounts of data, and making it easily accessible to the individual account holders, in the form of a chat interface.

Through these apps, chatbots are now able to provide retail banking customers their credit score, enable them to set and manage their budgets, and notify them about transactions from the convenience of the bank’s mobile application or website. More than retail banking, banking chatbots have started to assist customers in money management.

In this article we’ll review five banking chatbot applications, comparing their features and offerings. We’ll conclude with an overview of the future of banking chatbots.

(Readers with a more general interest in banking may want to read our full article on AI Applications at the Top 7 US Banks, and readers with a general interest in chatbots may be interested in our chatbot use-case analysis from last year.  Additionally, more information on how AI  applications can facilitate customer service,  download the Executive Brief for our AI in Banking Vendor Scorecard and Capability Map report. )

Comparing 5 Banking Chatbot Applications

Kasisto AI

  • Total funds raised: $28.5 million
  • Year founded: 2013
  • HQ location: New York
  • Number of employees: 11-50
  • Target user: Banking customers
  • Type(s) of data processed: Messaging

Kasisto‘s conversational AI platform, KAI is a banking chatbot and virtual assistant that can be deployed on a bank’s messaging, mobile and the web platforms. This tool is built with industry-specific knowledge to help customers with payments, transaction and account insights, and personal finance management.

Using natural language processing, AI reasoning and speech recognition technology, Kasisto claims that KAI is capable of intelligent, human-like conversations via text and voice, and can extract meaning and intent in communication. It can be deployed to multiple channels such as messaging, mobile and websites. In its website, the company claims that KAI can also be deployed on Internet of Things devices, but does not specify nor demonstrate which devices.

The 3-minute video below demonstrates some of KAI’s basic features:

According to the Kasisto website, the KAI application programming interface (API) is easy to use, and that banks can make the application available in a few weeks. It also includes a deep-learning analytical toolset for data collection and analysis, model training, testing, and deployment. Its self-serve customer portal also provides real-time reporting.

Kasisto’s claims to have major banks among its clients, including DBS in India, Singapore and Indonesia, Standard Chartered, Varo Money, Mastercard, and Royal Bank of Canada.

In a recent blog post, Kasisto CEO Zor Gorelev writes that KAI is DBS India’s first touchpoint with its customers. Since its launching under the bank’s own brand name, more than 1.8 million customers have joined digibank India. Today, the app  handles about 80 percent of customer inquiries and requests.

The executive team is led by Co-founder and CEO Zor Gorelov, who started as a developer at Microsoft. He later founded BuzzCompany.com, a provider of enterprise collaboration and messaging software, and SpeechCycle, a cloud-based contact center optimization solution for the telco market.

Co-founder and Chief Technology Officer Sasha Caskey has 15 years of experience in developing applications at IBM TJ Watson prior to Kasisto. He holds a Bachelor’s degree in computer science, minor in physics, and a Master’s degree in natural language processing.

Co-founder and Chief Product Officer Dror Oren, who served as executive director at SRI International, the research institute that is credited with what eventually became iPhone’s virtual assistant called Siri. He graduated with a degree in computer sciences and biology and holds a Master’s degree in business administration.

Personetics Assist

  • Total funds raised: $18 million
  • Year founded: 2010
  • HQ location: London
  • Number of employees: 51-100
  • Target user: Banking customers
  • Type(s) of data processed: Messaging

Personetics Technologies states that its Personetics Assist chatbot is a self-service solution built specifically for the financial services industry and used by its retail customers.

In the company website, Personetics Assist claims to be built on natural language processing, allowing it to comprehend what bank customers are asking for and respond appropriately. Personetics claims that itschatbot incorporates the customer’s most recent transaction data in the conversation, providing relevant responses based on the individual’s financial activity and banking relationship. Personetics Assist uses predictive analytics to anticipate customer questions and issues, and offer appropriate insights and advice.

From the bank’s website, mobile app, messaging platforms on Facebook Messenger and Amazon’s Alexa, Assist can also act on requests that require action such as sending payments, changing passwords and setting up appointments. The website shows that Assist is also being developed for Skype and other platforms.

We spoke with Nishant Chandra, Sr. Director of Data Products at VISA, about how Banking leaders can better handle the challenges that come with AI adoption in our podcast, AI in Banking. When asked about how AI is beginning to be integrated into banking tech stacks, Chandra stressed that it is more important to integrate AI capabilities into banking processes instead of simply layering them on top. He compares this type of integration to the layers in lasagna as opposed to the directly stacked pizza. Concerning this type of layering, Chandra said,

“Each layer of software when they are talking to each other are intelligent. They have data science capabilities built in, they have Ai intelligent ways of detect data fraud built in at every layer as opposed to doing it at the very end. These areintelligent software platforms which will transform or ingrain the AI capabilities in this space. This is a fundamental transformation that is happening.”

It is clear that Chandra finds it important that banks not let their AI applications get tacked on haphazardly at the end of whatever process they are adopted into. This may be even more important for banks or AI firms developing applications that themselves must be integrated into multiple channels such as Personetics Assist.

In a case study of Royal Bank of Canada, Personetics shows that since the launch of the product (rebranded as NOMI Find & Save) in October 2017, the bank garnered these benefits:

  • Client engagement with RBC mobile banking app increased by 20 percent
  • Average time-in-app increased 6 percent
  • More than 100 million insights read by clients in the first five months
  • Clients save 2x more regularly with NOMI Find & Save than they do with traditional savings products

Personetics claims that Assist is ready to deploy, cutting deployment time and maintenance effort for the banking clients.

To understand this app more, watch Personetics Co-Founder and Chief Executive Officer (CEO) David Sosna in this 2:14-minute video talk about cognitive banking at the Paris Fintech Forum in 2017:

Members of the executive leadership each have 15-20 years of experience in technology and financial technology services. It is not clear, however, if any of them have formal AI expertise.

Prior to Personetics, CEO David Sosna was also co-founder and CEO of Actimize, a provider of financial crime, risk and compliance solutions; and Gilon Information Systems, one of Israel’s largest business intelligence companies. Co-founder and Chief Operating Officer (COO) David Govrin’s focus lies in enterprise software solutions in the banking domain, specializing in business intelligence. He graduate from Tel Aviv University in Israel with B.Sc. and M.B.A. in Industrial Engineering and Management degrees.

Finn AI

  • Total funds raised: $3 million
  • Year founded: 2014
  • HQ location: Vancouver, British Columbia
  • Number of employees: 38
  • Target user: Banking customers
  • Type(s) of data processed: Messaging

Focusing only on AI for the banking industry, Finn AI claims to provide a personalized experience to banking customers through a conversational tool that uses natural language processing.

The company reports that its team of data scientists built machine learning processes into the app to recognize banking-related queries from customers. The company claims that through repeated interactions, it’s systems have improved to better recognize and reply to banking queries.

In the 6-minute Finovate Conference video below, Finn AI CEO Jake Tyler describes the major benefits of the product, along with a series of short demonstrations to walk through the app’s features::

To date, banks that have deployed Finn AI to clients such as the Bank of Montreal, Banpro, and ATB Financial.

Finn AI claims Bank of Montreal (BMO) as its first tier-one banking client in Canada. The bank rebranded the chatbot as BMO Bolt, which according to a Finn AI company statement, is capable of responding anytime of the day to 250 common questions relating to BMO products, foreign exchange rates, branch locations and ATMs.

The machine learning capability teaches the chatbot to learn responding to more questions from customers, allowing the technology to continually evolve. For more complex queries, the chatbot will customers to a live customer representative, also within Facebook Messenger.

In another press release, client ATB Financial claims to have made the app available to more than 1 million of its personal banking customers, who through the bank’s chatbot, can pay bills, view account balances, transfer money between accounts, as well as perform cross-currency money movement. ATB’s platform also offers access to spending insights, to a live customer representative, and Mastercard statement alerts.

Another client, Nicaragua-based bank Banpro, deployed a Spanish-language version of the chatbot, through which its customers can find a branch,  inquire about other bank products and services, view current account balances and recent transaction history within Facebook Messenger.

Following the clients’ deployment of the chatbots, however, no quantifiable results of the benefits have been reported.

Co-Founder and CEO Jake Tyler was a Director at Brook Intelligence and a Mergers & Acquisition Strategy Consultant at PMSI Consulting in London. He holds an MBA degree from the IE Business School in Spain. Dr. Kenneth Conroy, VP Data Science, focuses on the AI and deep machine learning areas of the product  and leads the development of the natural language processing system. He earned his Doctor’s degree in computer Science from the Dublin City University.

Clinc

  • Total funds raised: $59.8 million (edited December 2019)
  • Year founded: 2015
  • HQ location: Ann Arbor, Michigan
  • Number of employees: 11-50
  • Target user: Banking customers
  • Type(s) of data processed: Messaging, voice

Clinc offers several banking products targeted at individual customers. One is Finie Personal, which the company says can offer retail banking customers a personalized experience by providing detailed transaction-level financial responses and spending advice based on their banking and credit history.

The company website claims that the app understands complex questions and transforms them into useful insights for the use.

Fine Personal put together a 4-minute demo video to showcase its interface and some its features:

Another Clinc product is Finie Wealth, a chatbot for wealth management which the company claims has the intelligence to recommend relevant financial products. The company claims that this app can notify users about investment opportunities by providing stock prices and news, based on the user’s portfolio balance and performance, trading history and settlements.

The company also claims that banks can deploy both apps to a variety of channels and platforms including mobile, web, interactive voice response, messengers, and chatbots. To ensure security, Clinc’s engineers integrate the technology within the client’s premises. The complete suite comes with analytics, system administration tools, and a conversational AI training platform to help banking teams manage the back end.

Clinc’s executive team is led by CEO and Co-Founder Dr. Jason Mars who was recently recognized as one of the top 10 most innovative CEOs in Banking. He is a professor of computer science at the University of Michigan (UM, on leave) and conducts collaborative research with IBM Watson, Google, Facebook, Intel and the National Science Foundation. He has worked at Google and Intel and holds a Doctor of Philosophy degree in Computer Science. COO and Co-Founder Dr. Lingjia Tang is an assistant professor of Computer Science also at the UM. She has worked at Microsoft and Google, and received her PhD in Computer Science from the University of Virginia.

CTO and Co-Founder Dr. Michael Laurenzano received his PhD in Computer Science from The UM, with more than 40 research papers published. He also holds undergrad degrees in Mathematics and Computer Science, and graduate degrees in Computer Science and Engineering. Before Clinc, he worked as a computational scientist at University of California San Diego and as Director of Systems and Architecture at EP Analytics.

Trim

  • Total funds raised: $2.2 million
  • Year founded: 2015
  • HQ location: San Francisco, California
  • Number of employees: 14
  • Target user: banking customers
  • Type(s) of data processed: messaging

When Trim started out, it initially offered an app that the company claims can analyze and cut a consumer’s monthly spending by finding recurring subscriptions. The app then cancels the duplicate subscriptions, negotiates certain bills (Comcast, Time Warner, Charter), and  helps the consumer find better product or services.

If the user decides a subscription is no longer needed, Trim can be instructed to cancel it. The app is able to check a user’s recent transactions or see certain expenditures in the past month.

To view an independent review of how Trim works, watch this 4:33-minute video:

In the next five to 10 years, Trim executives claim the app will be able to manage basic financial decisions to help the user avoid credit card debt, too-expensive insurance, and late fees. The team also plans to build in functions that will help the user save up for retirement and achieve financial health.

The app is build on a platform that combines Google analytics, Heap infrastructure, and New Relic digital performance monitoring and management tools.

Trim’s executive team is led by Co-founder and CEO Thomas Smyth, a political science graduate of Yale University who forayed into venture capitalism for financial technology start-ups since 2014. He founded Trim  to focus on retail banking with the goal of improving Americans’ financial health.

Technical Co-Founder Daniel Petkevich is a mathematics specialist who led growth analytics and strategy at Redfin, co-founded Octane Lending, and was product manager for the Climate Corporation’s satellite imagery platform. He also graduated from Yale with a degree in Physics.

CTO Nick Fishman has been writing software for over a decade, previously worked at Google and Minerva Project, and has founded a number of startups.

Concluding Thoughts on Banking Chatbots

Banks are expanding their role beyond holding customers’ money to helping them manage their money. Financial technology companies are supporting this new role by developing AI-powered applications such as chatbots that increase customers’ financial literacy and well-being based on data-driven insights. While chatbots are unlikely to be among the top applications of AI in finance, they may be one of the few initial applications that direct impact the customer experience.

From a banking enterprise point of view, chatbots have the potential to support operations by automating customer support, providing 24×7 intelligent customer service, and possibly bringing in new customers or accounts. Chat apps may also be an effective way to push content to customers and analyze user engagement.

For retail banking customers, chatbots could offer a more convenient, personalized and enjoyable customer experience, giving users quick access to their banking and credit information at their fingertips. Banks have also begun to engage retail customers in discussions about spending patterns, investments, savings, and more to assist them in achieving financial health.

The future of Banking Chatbots

In the next five years, it is possible for banking customers to see chatbots improve the quality of conversation by answering more complex questions as their machine learning and natural language processing capabilities expand. It is also possible for chatbots to converse in more languages other than English as the startups expand their client base outside of the English-speaking countries.

Startups, however, face the dual challenges of finding talent with AI skills and organizational partners to fund development efforts.

For banks, the obstacle to be hurdled involves consumer trust in using new technology. In a study about the use of social media in banking published by the Queensland University of Technology in 2016, the authors discovered that while banking customers see perceived usefulness, and social and economic value in using social media, their sense of technology insecurity increased between 2010 and 2014, suggesting that this insecurity most likely stems from sharing financial information online.

In a separate survey, 84 percent of the bank executives surveyed believe that trust is the pillar of the digital economy. To successfully adopt AI, however, banks need to strengthen their security strategies to increase customers’ digital trust.

Some global banks with deep resources have already adopted  chatbots given its operational and analytics benefits. For smaller banks, however, it may be wise to initially pilot the technology in certain markets until the technology becomes more robust, security more stable, and trust is unquestionable.

 

Header image credit: Green America