• Erica was launched in 2018, developed by Michelle Moore, the former head of digital banking at Bank of America (BoA).
• It was initially launched to help consumers adopt better money management habits. Today, it provides a more intuitive and personalized customer service experience using AI technology.
• In July 2023, Erica surpassed 1.5 billion client interactions, totaling nearly 10 million hours of conversation. During this time, it assisted nearly 37 million BoA clients with their everyday financial needs.
• The effectiveness of helpful answers was reported to be 90% in 2021.

AI chatbots and conversational chatbots have emerged as pivotal tools for customer engagement strategies.

Over 98 million users (approximately 37% of the US population) have engaged with a bank’s chatbot in 2022. This number is projected to reach a staggering 110.9 million users by 2026. Also, as per research by the Consumer Financial Protection Bureau (CFPB), the top 10 commercial banks in the US have already deployed chatbots as a part of their customer service.

When Bank of America (BoA) introduced a state-of-the art AI chatbot, Erica, it set a new benchmark for operational efficiency in the highly competitive sector of banking and finance.

Survey shows the number of individuals aged 18+ who have interacted with their bank via an AI-enabled chatbot to obtain information or receive customer service at least once during the calendar year, source

Traditional v/s modern banking: the problem that boa wanted to solve

Essentially, Bank of America was looking to foster a more interactive, responsive, and personalized banking environment. Their intention was to bridge the service gap between traditional banking hours and the modern customer's expectation for on-demand, immediate service.

Briefly, they were looking to solve these two problems:

• Accessibility to banking services: Traditional banking hours and human-operated customer service lines often result in extended waiting times. To alleviate this constrained accessibility, Bank of America sought to significantly enhance its customer engagement strategy while also simplifying basic banking transactions.
• Democratize financial advice: As a market leader in both mobile banking and AI implementations within the US, they were also looking to democratize personalized financial guidance, which was typically reserved for their top-tier customers.

The solution: what is erica?

With Erica, the company hopes to help consumers create better money habits - Michelle Moore, Bank of America’s Head of Digital Banking, 2016.

In 2018, Bank of America assembled a dedicated team of over 100 individuals with a singular mission: to build Erica, a chatbot envisioned to assist customers in navigating their financial lives. Erica is an AI-driven chatbot designed to enhance the banking experience for BoA mobile app users. Through a variety of features and interaction modes, Erica has heralded a new level of convenience and accessibility in digital banking for BoA mobile app users.

Features and unique functionalities of erica:

• Interaction modes: Users can communicate with Erica to conduct transactions via voice commands, text, or a simple tap on the mobile screen.
• Access to transactions: Customers can quickly access balance information, make money transfers, request account/routing information, pay bills, view insights, etc.
• Personalized insights: Erica analyzes the transaction data and provides report to help users scrutinize their finances, identify any increased subscription fees, indicate any duplicate transactions, and even monitor the receipt of merchant refunds.
• Proactive information: This feature was introduced in the first half of 2023, where Erica could connect users to financial specialists when they had specific questions around BoA’s new products and services.

Technology behind erica

Central to Erica's capabilities is Natural Language Processing (NLP), enabling it to engage in human-like dialogues and foster intuitive interactions with customers. Complementing this is the application of predictive analytics, which allows Erica to provide personalized financial insights by analyzing users' transaction history and financial behavior.

BoA fed Erica with 10 million hours of customer interactions so that it would be trained to answer the most frequently asked questions by customers. With this data, along with thousands of other Machine Learning (ML) inputs, Bank Of America knew almost exactly what queries to program into the algorithm behind Erica. Finally, to ensure security, Erica operates within a stringent security framework, safeguarding sensitive financial information while facilitating secure engagements.

So, what was the impact of Erica?

Erica has reportedly powered a 19% spike in BoA earnings to reach $7.41 billion in Q2 2023.

For BoA, Erica marked a significant milestone in harnessing AI technology to enhance customer engagement and streamline banking operations. The impact of Erica can be seen through multiple lenses:

• User adoption and engagement: Within a year of its launch, Erica had already amassed 6.3 million users and facilitated 16.5 million interactions. In 2023, users had engaged with Erica more than 333 million times, up 35% year-over-year.
• Customer engagement even through crisis: During the pandemic, it learned over 60,000 coronavirus-related terms to process a range of inquiries related to COVID-19. This demonstrated the system’s agility in adapting to emerging customer needs and global developments.
• Operational efficiency: Its ability to autonomously handle banking inquiries significantly reduced the load on human customer service representatives, thereby improving operational efficacy.

What can you learn from boa’s approach towards erica?

In 2022, BoA decided to add a human element to Erica.

Even though Erica was actively used across all age groups, BoA realized that consumers still desired human interaction. This is supported by a study where 54% of consumers mentioned that they’re more likely to interact with a human when it comes to sensitive financial information. Therefore, in a bid to enhance their AI capabilities without compromising on human interaction, BoA decided to go for a hybrid approach. They upgraded Erica to seamlessly shift the conversation between a human agent and the AI.

Along with the above, below are three more things that boa did right while developing erica:

• Training with the right data: Erica was able to predict popular questions that customers might ask because of the extensive customer data that BoA fed the bot.
• Iterative design process: Driven by machine learning algorithms, Erica can continuously learn and improve. This allows it to regularly refine itself based on user interactions, ensuring relevant and hyper-personalized services.
• User-centric design: Aiming to simplify banking interactions, Erica's knowledge base and training have a user-centric approach. With its adaptive interaction capabilities, it can respond to users' preferred communication modes and offer solutions ranging from financial insights to subscription management, enhancing the user experience.

Have you made any financial transactions using an AI chatbot with your bank? Let us know your thoughts on Erica and the usage of AI chatbots in the financial sector.

This article was first published here.