Conversational Interfaces in Banking and Financial Services

The banking and financial services sector is changing rapidly. Emergence of Fintech startups, consumer movement towards digital channels, increasing computer power, and enhanced technology, are creating opportunities for innovative solutions.

Big (of ‘BigTech’) companies such as Amazon, Google and Facebook establish the precedents for new technologies. Consumers now expect to receive personalised service through multiple channels: online, mobile, telephonic, and so on. In 2020, startup banks like Monzo, Revolut, and N26 are emerging as leaders in the European digital market, while “neobanks” are appearing across the US to attract younger customers not looking for physical branches.

Putting everything together, banking customers now have more digital, high-touch and simplified experiences working with their banks compared to the branch-led operations of old. Consumers who use a mix of channels, outside of in-person, want personalised experiences from their financial institutions. To meet this rising demand, firms must engage with customers in the right places at the right times. Industry incumbents are turning to artificially intelligent conversational interfaces, known as chatbots, to improve customer experiences while reducing costs.


What is a chatbot?

Conversational interfaces tend to be known as chatbots or bots. They are computers or machines simulating human conversations, either via text or voice. The most popular chatbots in recent years are Alexa, Siri, and Google Home, which are all voice-activated.

A chatbot works by taking the words and phrases of a user, processing them through algorithms, and returning a pre-set answer using a series of rules. Businesses will use platforms like Facebook Messenger, WhatsApp, Slack, or Skype to create bot programs that are familiar to users. Today, there are three different types of chatbots.


A rule-based chatbot is the simplest form of technology. Users can click pre-defined options with relevant answers returned, depending on the selection. They are best for basic scenarios that do not have a vast number of potential responses. An example is below:


Intellectually independent chatbots

An intellectually independent chatbot will learn from user inputs and requests. The method is based on an application of artificial intelligence (machine learning), which allows a computer to learn from data by recognizing patterns. The chatbot will train itself to interpret questions over time, as users enter more data into the program. Essentially, they will start to create their own rules.

AI-powered chatbots

An artificially intelligent chatbot combines rules and data to simulate human intelligence. The best chatbots can understand the context of free language, but will also be set up with predefined flows, ensuring they solve a problem. Like humans, chatbots switch between conversations and scenarios to address user requests in the right contexts.

The algorithms powering conversational chatbots use natural language processing (NLP) to make interactions feel like genuine human-to-human experiences.


Chatbots are no longer those programs struggling to answer questions and leaving customers in limbo. They are now advanced enough to handle entire conversations, acting as smart digital assistants that personalise experiences just as well, if not better, than human interactions.

According to The Financial Brand, AI will save the banking industry over $1 trillion by 2030, with chatbots playing a decisive role. A study by Juniper predicts that the use of chatbots will save banks up to $7.3 billion worldwide by 2023.

They reduce the need for customers to speak with human agents, are available 24/7, learn independently, and offer the experiences 21st-century consumers demand. To put this into perspective, chatbots can save 862 million hours of work.


How chatbots are being used in financial services

An increasing number of financial institutions are turning to chatbot technology to enhance their operations and customer service. The main benefits are:

  1. Reducing costs: compared to hiring, training, and retaining human workers, chatbots are inexpensive. They need coding, data, and storage, which is all coming down in price as demand continues to increase.
  2. Quick and easy: customers can use chatbots via mobile, desktop, or within apps for fast access without picking up the phone or formalising an email
  3. Conversational: chatbots are more personal than emails or forms. They can use data to create a real-time customer experience.
  4. Financial advice: chatbots access data quickly and provide customers with financial advice and recommendations. They can analyse spending habits faster than the human brain.
  5. 24/7 support: chatbots are available whenever customers need them, offering dedicated support.

There are plenty of financial firms making the best use out of chatbot technology.

JPMorgan Chase

JPMorgan Chase uses COIN to analyse complex contracts more quickly than human lawyers. JPMorgan states that the bot saves more than 360,000 hours of human labour. Since launching the chatbot in 2016 to improve back-office operations, JPMorgan Chase reviews contracts with minimal errors within seconds.

Bank of America

Bank of America is a market leader when it comes to mobile banking and the use of AI. The AI-powered Erica chatbot is like a personal banker. Using machine learning algorithms to process data rapidly, Erica learns from transactions to assist customers. Erica provides balance information, recommends how to save money, issues credit report updates, and performs simple transactions.

Erica is incrementally expanding its capabilities through data, becoming more intelligent. Customers can interact with Erica using either voice or text.

HSBC (Hong Kong)

HSBC Hong Kong offers a virtual assistant chatbot for corporate banking known as Amy. Customers receive 24/7 support in several languages about banking products. Amy learns through a customer feedback mechanism, absorbing data from each and every experience.



The use of a female chatbot by HSBC is a deliberate move to dispel stereotypes in the sector.

SEB (Sweden)

In early 2017, the Swedish bank rolled out Aida to act as a trainee for the bank’s front end customer service. Aida can handle IT support questions, help customers with card issues or account queries, and book meetings. Its skills are always expanding with experience, with even the ability to sense emotions (though there is some ways to go before reaching a human’s level).

Commonwealth Bank (Australia)

The Commonwealth Bank of Australia uses Ceba for up to 200 tasks, including: card activations, balance checks, payments and queries. The bot understands over 500,000 ways customers can ask about different activities, allowing it to be genuinely conversational.


Limitations of chatbots

With every new technology comes barriers to entry. Chatbots do tend to have some inherent problems which industry incumbents need to be aware of before they embark on further projects.

Firstly, chatbots can struggle with languages and dialects. It is challenging to find enough data to account for a multicultural world. It will take much trial and error to get it right.

Consumers are not all yet comfortable with chatbots for financial queries. Many people prefer to speak with a human as they don’t trust a robot can offer the same level of service. There needs to be further awareness, globally, on what chatbots offer.

New technology also comes with security risks. Banking is a vulnerable sector for cybercrime, given the data firms hold. Chatbot developers need to provide clear guidance on the security measures in place if they expect customers to share their information.



While chatbots are still limited, they will become increasingly valuable as they continue to organically self-learn. They provide opportunities for financial organisations to create personalized experiences that reduce operational costs and improve efficiencies.

The use of chatbots in banking and finance is still in its early stages, and many are still exploring the potential. If they counter the security risks through new technologies like facial recognition, blockchain, and IoT, the bank of 2030 will look incredibly different to what we have today.

Disclaimer:  The author of this text, Jean Chalopin, is a global business leader with a background encompassing banking, biotech and entertainment. Mr. Chalopin is Chairman of Deltec International Group,

The co-author of this text, Robin Trehan, has a bachelor’s degree in economics, a master’s in international business and finance, and an MBA in electronic business. Mr. Trehan is a Senior VP at Deltec International Group,

The views, thoughts, and opinions expressed in this text are solely the views of the authors, and do not necessarily reflect those of Deltec International Group, its subsidiaries, and/or its employees.