Support customers through chatbots
Introduction
Chatbots are becoming increasingly popular in the banking and finance industry as a way to provide efficient and personalized customer service. These AI-powered assistants are available 24/7 and can answer a wide range of questions, from account balances to loan applications. By automating routine tasks and providing instant responses, chatbots can free up human customer service representatives to focus on more complex issues. Additionally, chatbots can be customized to suit the specific needs of the banking and finance industry, making them an ideal solution for financial institutions looking to improve customer service and increase efficiency. In this page we will discuss about the development of a prototype Contextual AI assistant for a banking and finance use case.
Use Case
In our use case, we developed a prototype Contextual AI assistant (chatbot) using the Rasa framework. The bot was designed to assist customers with questions about financial products and services offered by our client, a financial institution. The chatbot was trained on a variety of financial products such as savings accounts, loans, credit cards, and investment options. The bot was able to answer questions about interest rates, fees, and other product details. Additionally, the chatbot was programmed to upsell and cross-sell financial products to customers based on their needs and preferences. For example, if a customer was inquiring about a savings account, the chatbot would also suggest other products such as a credit card or loan that may be of interest. This helped to increase the chances of customers purchasing multiple products from the financial institution.
Technical Implementation
The chatbot was developed using the Rasa framework, an open-source machine learning platform for building conversational assistants. Rasa allows chatbots easy creation and customization through its natural language processing (NLP) and machine learning capabilities. The bot was built using Rasa's core library and the Rasa NLU library for natural language understanding. Additionally, the bot was integrated with a custom action server to perform specific tasks such as fetching product information from the financial institution's database. The prototype was also integrated with a messenger platform such as Facebook messenger, WhatsApp, or Telegram to allow customers to interact with the chatbot through their preferred messaging channel. The chatbot's performance was tracked and monitored using Rasa's analytics and monitoring capabilities.
To train the chatbot, we used a combination of pre-existing and custom data we created. The custom data was created by our team, by writing the interactions between a customer and the chatbot, this way we were able to train the bot to understand the financial domain. The chatbot was also fine-tuned with real-user interactions to improve its performance.
The chatbot was also designed to handle customers’ requests and questions in a conversational way, by using context and memory, this way the customer would feel like talking to a human.
The technical implementation of the chatbot was designed to be flexible and scalable, allowing for easy integration with other systems and future updates.
Results
The chatbot was deployed in a pilot phase for a limited number of customers. During this period, we collected data on customer interactions with the bot, including metrics such as customer satisfaction, sales conversions, and the average time it took to resolve customer inquiries.
The results of the pilot phase were positive, with customer satisfaction rates at 80%. This means that 80% of the customers were satisfied with the assistance provided by the chatbot. Additionally, the chatbot converted 20% of customer inquiries into sales, which is a significant increase from the financial institution's previous conversion rate. The average time it took for the bot to resolve customer inquiries was less than 2 minutes, providing a fast and efficient service.
We also noticed that the chatbot was able to handle a large number of customer interactions simultaneously, without the need for additional human support, which results in cost savings for the financial institution.
The chatbot implementation successfully provided efficient, personalized service to customers and increased sales conversions for the financial institution.
Conclusion
In this page, we discussed the use of chatbots in the banking and finance industry, specifically the development of a prototype Contextual AI assistant (chatbot) using the Rasa framework. The bot was designed to assist customers with questions about financial products and services offered by a financial institution, with the ability to upsell and cross-sell products. The chatbot was developed using the Rasa framework and integrated with a messenger platform to allow for easy customer interactions.
The results of the pilot phase were positive, with customer satisfaction rates at 80% and a conversion rate of 20%, resulting in significant cost savings for the financial institution. This demonstrates the potential for chatbots to provide efficient and personalized service to customers in the banking and finance industry.
If you're interested in implementing a similar solution for your financial institution or you want to know more about this use case, please don't hesitate to contact us. Our team of experts can help you understand how chatbots can help your business and walk you through the implementation process.
Want to learn more?
Miguel Cabrita
Senior Data Scientist, Co-founder
Miguel has helped various companies in banking and finance implement lead scoring and AI solutions.
Having a strong technological background and understanding of business processes and the banking industry helps him detect specific needs and offer the necessary AI solutions for each of them.
