7 Chatbots in the Financial Industry – Paypal, Kasisto, and More

Niccolo Mejia

Niccolo is a content writer and Junior Analyst at Emerj, developing both web content and helping with quantitative research. He holds a bachelor's degree in Writing, Literature, and Publishing from Emerson College.

7 Chatbots in the Financial Industry - Paypal, Kasisto, and More

Many financial institutions are experimenting with chatbots both for general customer service and for offering new and better financial services to their customers. In addition to banks and insurance companies, other types of financial services companies can benefit from this type of application as well. Financial customers can now check the status of their loan applications and stock portfolios and request refunds using AI-powered conversational interfaces.

In this article, we list seven chatbots available to customers in banking and personal finance, with specific use-cases such as:

  • Banking and account services 
  • Customer service agent assistance 
  • Voice-enabled virtual assistants
  • Requesting and managing refunds 
  • Flexible chatbot platforms for multiple uses 
  • Self-service insurance transactions

Our list of financial services chatbots begins with Kasisto, an AI vendor that ranked among the top customer service companies in our AI in Banking Vendor Scorecard and Capability Map report.

1. Kasisto 

Kasisto has an AI platform called KAI that they claim can help financial institutions create chatbots for their customers to ask questions and use to make payments and review account details. The company also claims KAI chatbots can help users manage their funds from multiple accounts. The software can send a chatbot’s conversation to a human customer service employee in cases where it cannot solve a customer’s problem.

Chatbots developed using Kasisto’s platform can integrate into smartphone apps, websites, and dashboards for backend employees. The company also claims to have a deep learning tool for business banking chatbots that helps train new machine learning models, however, they do not offer a detailed explanation of how multiple neural networks would work to make a chatbot more effective

These chatbots can purportedly converse with customers about financial tasks such as applying for credit cards, product discovery and managing funds. They can also fulfill some requests directly within the chatbot interface. These include sending or scheduling payments or updating the terms of a claim.

They can also purportedly fulfill these requests within the digital channel in which they are installed. For example, a chatbot may be able to direct the customer to a page where they can confirm a payment and then autofill their identification information such as their name. Additionally, this system can prompt customers with recommendations on more efficient ways to complete recent actions.

One example of this is a prompt that reads, “You Sent 100 US dollar wires of $50 each to Singapore Yesterday. Did you know you could send a foreign-exchange ACH payment instead? Sign up here.”

The video below is a demonstration from Kasisto showing a Mastercard chatbot made on their platform. It shows how the chatbot can handle customer questions about personal data such as account balances:

2. Digital Genius

Digital Genius claims their chatbot software “Co-Pilot” helps businesses automate the most frequently asked customer support questions. The software can also integrate with other management tools such as Salesforce and Zendesk, which could assist customer service agents in finding answers to customer questions. Natural language processing (NLP) could make this possible because the chatbot would be able to parse the words within a customer’s question.

The Co-Pilot chatbot can purportedly find possible answers to customer questions automatically, and these answers are based on both historical customer service data. It can then send that response to the customer or withhold the response for a human agent to approve.

This could result in the chatbot making more nuanced responses as it continues to adapt from approved or disapproved responses.

Digital Genius highlights confidence as an important value point in their product. Their apps use confidence intervals to determine how accurate a customer service answer might be, and then hold back low-scoring answers for approval by a human agent. They claim this results in higher accuracy before the service agent tells the customer the answer.

The following video is a demo of Digital Genius’ chatbot from their president and co-founder Mikhail Naumov. The demonstration occurs at 0:00 and ends at 3:00:

3. Nuance Virtual Assistant

Nuance Communications is more widely known for their AI-enabled voice recognition technology for healthcare. However, more recently, they have begun to offer a virtual assistant for customer service called Nina. Users can ask the virtual assistant questions with voice or text, and the company claims it can be integrated into a client company’s website or smartphone app.

This software is built for customer service for larger companies, and so it is also made to be compatible with SMS texting applications and smart TVs.

One example of how this could benefit financial services companies is in customer onboarding and advisory. A new customer could have difficulty setting up their new account, but would have access to a virtual assistant that could effectively walk them through it by asking for each piece of information individually.

Nuance’s Nina virtual assistant can also be set to display automatic prompts along the lines of certain business rules. It could pop up and offer help if a customer is spending a long time on the same few financial plan options. This may require multiple subsequent questions for clarity, but the assistant would soon be able to navigate the user to more information that might help them make a decision.

Nina can also be integrated into human-assisted customer service engagements to help service agents. Customer service applications typically switch to a live chat with a human customer service agent if the chatbot cannot recognize the question. The application can purportedly consolidate all information regarding the customer service ticket with a hidden, more experienced agent acting as a “coach.” 

This way a customer service agent can get assistance from their teammates faster. Additionally, Nina can take that consolidated information and update a service agent with it when it switches to live chat.

The following video is a labeled demonstration of how Nuance’s virtual assistant works for solving customer service issues. Here, the chatbot intakes typed questions from a smartphone:

4. Paypal

Paypal also offers an AI-powered customer service chatbot, and it can run through Facebook messenger. The chatbot asks users to log in each time they use it to access personal data for security reasons. It can also bring up a list of the customer’s disputed payments so they may make sure to check the status of each one.

Users can ask the chatbot questions about nearly any aspect of their PayPal account. Common topics include:

  • Declined payments: The reason why the customer’s payment cannot clear or why they cannot accept money that has been sent to them.
  • Unauthorized charges: Inquiring about payments for goods or services that the customer did not purchase, which is followed by how to rectify them.
  • Passwords: This also includes other login information such as the email linked to the account and security questions.
  • Account Holds and Limitations: Any condition set on the user’s account that limits how much money they can send or accept, and how frequently they are allowed to.

This application also covers refunds and can help the user check the status of a refund or request a new one. Users can also check the processing time of a given payment or refund, which may put their mind at ease while waiting for it to clear.

Paypal also claims to use ML for fraud detection and risk mitigation in their payments platform. They purport to use hundreds of identifying factors within each transaction to ensure no fraudulent activity can go on. This may also include chatbot interactions such as asking certain questions repeatedly within a certain timeframe. While there are likely an equal amount of minor identifying factors within payments and chatbot conversations, it is unclear whether these conversations are also used to track fraud.

5. Passage AI

Passage AI has an AI platform for creating conversational interfaces for financial services companies. They claim their platform can help these financial services companies with giving their customers easy access to their bank accounts, investment funds, or credit card information. 

Additionally, they purport to be able to help these customers find home loans. The platform also has the capability to update users on the status of their loan or credit card application. A chatbot made with this platform can also answer more simple customer service questions such as locating a routing number.

Passage AI claims that they have multiple machine learning models which have been prepared for various business areas in banking. These machine learning models have purportedly been pre-trained on enterprise data for each of these use cases. The business areas are as follows: 

  • Credit Information: Disclosure, Freezing one’s credit, and creating credit reports
  • Credit Cards: Balance inquiries, bill inquiries, and credit line increases
  • Applications: Applying and checking the status of applications for credit cards and loans
  • Paying off Loans: Making and scheduling payments
  • Payroll and Invoicing: Small businesses can use this interface for their employee’s paychecks
  • Retirement planning: Using individual account information to make a personalized retirement plan

Using Passage AI’s platform, a financial institution may be able to create a chatbot that could do any number of these things. Alternatively, they could focus on one or two specific capabilities and focus machine learning training in that direction. Depending on the needs of the client, Passage AI’s platform could help build a chatbot that can source and translate the most important information into a suitable response for the user.

The following video shows how Passage chatbots can be used beyond the listed capabilities and assist customers in shopping:

6. Kinvey Native Chat

Kinvey Native Chat is Progress Software’s chatbot platform that they claim helps insurance companies build chatbots for customer self-service transactions. Common uses of these chatbots include selecting insurance policies and scheduling appointments. This enables the user to make appointments and purchase insurance without speaking to a human employee, which can save time for the client insurance company.

In order to create a chatbot that can process entire transactions, Kinvey and their clients need to work to train it in a way that will allow it to offer the user certain transactions and them process them accurately.

The software likely categorizes inquiries as a call for help, a question, or a request to file a claim. This type of chatbot would operate as described when the inquiries are scored above a pre-established confidence threshold. Should an inquiry have a confidence level below that threshold, the chatbot can route the question to a human employee for review.

The video below demonstrates how a chatbot made with Kinvey Native Chat helps a user find the best insurance policy for them:

7. Finn AI

Finn AI has a banking chatbot service which can also be enabled for multiple languages and sentiment analysis that allows the client company to detect the quality of the customer’s experience. Because of this, the chatbot has the ability to determine whether it gave a helpful response based on how the customer responds to it. This could involve managing funds, payments, loans, or applications. 

Financial institutions may benefit from this type of sentiment analysis because they can see trends in the types of customer service questions they receive. This could influence businesses to update their readily available information to fit those needs. They can also more closely monitor the chatbot’s gradual improvement over time.

In order to handle multiple languages, Finn AI’s chatbot would require a machine learning model that was trained to recognize topics and phrases out of multiple sets of words at a time.

These sets would correspond to individual languages, and the data science team working on the model would have to make sure it can discern when a word is used in different languages.

The software can likely correlate phrases and individual words from any language it is programmed to understand into helpful data points.For example, a Finn AI enabled chatbot made for English and Spanish could then respond to each instance of a question in either language with an answer that holds the same meaning across both.

Header Image Credit: Nord VPN

 

Subscribe
subscribe-image
Stay Ahead of the Machine Learning Curve

Join over 20,000 AI-focused business leaders and receive our latest AI research and trends delivered weekly.

Thanks for subscribing to the Emerj "AI Advantage" newsletter, check your email inbox for confirmation.