Big banks were first to market with financial chatbots powered by conversational AI. In North America, this included Capital One with the debut of Eno in March 2017, followed by Bank of America’s Erica, launching in June 2018. Since these early days, we’ve seen a strong adoption rate with many predicting that most retail banks will have some form of chatbot or virtual assistant by mid-2021.  

 

At major banks these projects have moved well beyond innovation experiments. According to Bank of America, as of December 2019, Erica has surpassed 10 million users and is on track to complete 100 million client requests by early 2020. At a time when branch and phone channels are increasingly under pressure as a result of the fall out from Covid-19 a new channel that instantly handles millions of customer interactions, 24/7 is a powerful tool. 

 

 

This is an impressive achievement and Erica has quickly set the standard for banks and credit unions to follow. Whether or not it’s a realistic model that you can replicate comes down to cost, which was (and is) significant. Published reports indicate that Bank of America had to amass a team of 100+ people and over two years of effort to bring Erica to market – at an estimated cost of about $30 million. As Erica continues to scale, so will the investment. 

 

Let’s face it, there are few financial institutions globally that can fund this type of project at this scale. But as virtual assistants become widely adopted amongst the largest FIs, regional banks and credit unions will need to assess how they can deliver their own take on this experience, without building a 100+ person team of data scientists to manage it. 

 

Adoption of AI-Powered Banking Chatbots

The reality is that AI-powered chatbots are being adopted by most industries at a phenomenal pace. Leading the charge is the banking and insurance sector. 

 

 

The experts at Gartner predict that by 2022, 70 percent of all customer interactions will involve emerging tools like chatbots, machine learning, and mobile messaging – up from 15 percent in 2018. So how can mid-sized banks and credit unions hope to compete when Bank of America has set the bar so high?

 

Resourcing is Critical

The challenge for regional banks and credit unions is how to deliver a similar experience and gain the same operating efficiencies without a team of 100+ data scientists and engineers.

Two to three years ago, off-the-shelf NLP and chatbot building tools were viewed as the panacea. However, as many banks and credit unions have learned, these solutions resolve only a small part of the problem. Plus they still require a significant investment (10+ FTEs) to operate in real-world production. This has led to many projects failing, with the experts at Gartner predicting that 40% of chatbot projects started in 2018 will have failed by the end of 2020.

 

In the conversations we’ve had with many banks and credit unions, the frustration is evident. As a result, we’ve focused our product development to capture the end user and business value of AI-powered chatbots, without the long set-up times and high operating costs.

 

We help our customers right-size their resource strategy. We relieve them of the need to support the science and technology, and instead, we do this, providing best-in-class language understanding, a dedicated library of banking-specific utterances, and predefined responses. Rather than investing two or more years pre-launch, they can leverage a virtual assistant that is already trained in banking, out of the box.

 

Gartner predicts 40% of chatbot projects started in 2018 will have failed by end of 2020. 

This allows our customers to focus on the deliverables that are unique to their organization, including: 

  • Unique content (products, services, etc.)
  • Integration of existing internal systems (we integrate seamlessly with banking technology for fast deployment times)
  • Analytics to measure success and areas of improvement
  • Alert criteria to ensure important results are flagged 

Pre-Trained Banking Chatbots

The Finn AI dedicated library of utterances and responses is a significant differentiator. It allows mid-sized organizations to operate at a scale well beyond their existing customer base. 

 

Compiling all the different ways people will converse with your chatbot is incredibly convoluted and time-consuming. Even the former head of digital banking at Bank of America, Michelle Moore, was surprised to discover that customers were able to find 2,000 different ways to ask Erica to send money. Now extrapolate this across all the different requests your chatbot will receive to better understand the complexity. 

 

Our data aggregation model provides you with the ability to support over 80 percent of the bank and credit union queries you’ll receive at launch—and you won’t need two years to get there.

 

High Maintenance Costs 

Although Erica launched with the support of 100+ experts and a $30 million budget, it’s a given that these numbers have continued to increase as more conversations and capabilities are introduced.  

 


Conversational AI needs to be continually refined and refactored to ensure it’s working optimally. Few of our customers can provide the resources required to carry their deployment into the future. Finn AI provides the necessary support and ongoing maintenance to ensure your virtual assistant is on point and functioning properly as your chatbot scales. 

Build Versus Buy

Building an AI-powered virtual assistant is possible, but it’s also ridiculously expensive. Bank of America, with 66 million consumer and small business clients, has the scale and resources to do this in-house. But thousands of banks and credit unions do not. 

 

Almost two years after the launch of Erica, no one at Bank of America, nor any analysts or industry leaders are willing to admit Erica is a success. While the jury remains out on Erica, the adoption of chatbots within the financial sector continues to surge ahead. 

 

You have the opportunity to embrace this evolution of conversational AI in banking, as long as your strategy is realistic and aligns with the budget and resources you have at your disposal.

 

Do you already use live-chat or messaging for support? 

Take advantage of our free Call Driver Analysis. We will review your conversation logs and provide you with a detailed analysis of what your customers are asking, free of charge. The analysis will allow you to clearly define the business impact a chatbot can create for your bank or credit union. 

 

Gartner Report AI

 

 

Access our intelligent chatbot from your live-chat platform to automate routine tasks, 24/7 from the comfort of your own home. Contact us.