Thanks to ChatGPT, 2023 is the year of the chatbot in banking

OBSERVATIONS FROM THE FINTECH SNARK TANK

Among my 5 predictions for banking and fintech in 2023, I wrote:

“2023 will be the ‘year of the chatbot’ in banking. After years of hearing pundits and futurists tell them how disruptive AI is going to be in banking, 2023 will finally be the year bank leaders do something about it.”

It wasn’t a well-received prediction, but I’m sticking to my guns (not literally, of course).

Three key reasons why 2023 will be the year of the chatbot – or conversational AI, more broadly – ​​include: 1) The need to improve digital service; 2) The need to improve the employee experience; and 3) Banks will experiment with ChatBPT.

Digital service needs (developed) chatbots

I know what you’re thinking: “Dude, what are you smoking? Haven’t used chatbots? The experience is terrible!”

Consumer research refutes this view. According to a study by Cornerstone Advisors, consumer ratings of the mobile banking experience are higher for banks with a digital assistant than for those without one.

But not all “chatbots” are created equal.

Chatbots are “evolving” to become intelligent digital assistants (IDAs).

Although the terms chatbots and IDAs are often used interchangeably, according to a report by Cornerstone Advisors titled The chatbot journey: Turn intelligent digital assistants into integral members of the teamthere are differences:

“Chatbots are typically rule-based systems that can perform routine tasks with general FAQs. IDAs are fully equipped with natural language understanding that helps understand and retain context for polished conversations while performing a variety of tasks to meet a user’s requirements.”

Intelligent digital assistants provide superior service by:

• Being conversational. Basic (ie non-developed) chatbots draw from a limited library of scripts and FAQs. This approach only provides an easy path to a predetermined answer. IDAs, on the other hand, are pre-trained with knowledge from the customers’ financial history and behavioral patterns, which gives them a more extensive conversational base of experiences, language and terms to draw on to meet specific customer needs.
• Counseling versus solving. The primary role of banking chatbots is to quickly resolve consumers’ basic transaction questions, or move them to human intervention. This limitation often leads to incomplete problem resolution and a high customer churn rate. IDAs, on the other hand, act as knowledgeable bankers who can walk with a client and recommend the most informed next step in their specific financial journeys.

Chatbots fill support gaps without much capacity to retain and develop relationships. Using conversational skills, a deep data library and AI-powered analysis of usage patterns, IDAs understand what customers are asking for and can guide them to what they want while encouraging them to explore other engagement options.

Conversational AI improves the employee experience

It’s true that many consumers (today) will resist using chatbots, preferring to deal with another human. But have you been to a bank branch or called a bank’s contact center recently?

Eight out of ten banks struggle to recruit new employees, according to Cornerstone Advisors. When these banks find someone to come on board, it takes a long time to get them up to date on products and processes.

The new reality: Chatbots are for employees – and are the new employees.

Employees often turn to other employees for help figuring out how to respond to customer inquiries, but what do they do when their colleagues don’t have the answers?

Banks are increasingly using conversational AI technology to support staff directly – effectively turning a chatbot into a ‘member of the team’.

Making a chatbot or intelligent digital assistant a member of the team is like bringing a new human employee into the team. If you hire someone (a real person) into your organization – what would you do to ensure that person succeeds?

You’ll create an onboarding plan, assign that person to report to one of your top managers, and create a professional development plan with a multi-year time frame to identify the types of roles and positions you want the person to fill on his or her way to management level.

It is no different for a chatbot on its journey to become an intelligent digital assistant.

Banks will experiment with ChatGPT

Bank and credit union CEOs who don’t instruct their CIOs and CTOs to report back to the executive team with ideas for how they can ChatGPT are derelict.

The recently announced conversational AI tool from OpenAI is great for composing poems in the vein of Post Malone, but there are more mundane uses for the tool in banking. In a recent LinkedIn post, Chris Nichols, director of capital markets at SouthState Bank, identified 15 use cases for ChatGPT in banking. My favorites included:

• Create code. ChatGPT can analyze all open source code and synthesize code libraries to create code capsules. Programmers at SouthState have asked ChatGPT to: 1 Write python code to create a graph of the current month’s expenses; 2) Write C+ code that will match an email address with the one registered; and 3) Write Java code to create a poll for the bank’s website.
• Product design. Nichols points out that one of ChatGPT’s capabilities is taking on a specific customer persona, such as a doctor, retiree, CEO or engineer. ChatGPT can tell a bank: 1) How to pitch treasury management services to a Controller in a municipality, and 2) How a lawyer wants to be notified that the bank has placed a freeze on their checking account.
• Legal contracts. Chat GPT may not be ready to write and analyze legal contracts, Nichols says “it’s almost there,” saying his bank uses the tool to “put in missing clauses about return of information, venue, non-automatic renewal, regulatory requests , and other elements of draft contracts” saves the legal team considerable time.

Conversational AI is a fundamental technology in banking

Conversational AI, has become a competitive necessity – i.e. a foundational technology – not only to provide customer and employee support, but because of the need to collect data.

Attempts to codify and store “data” collected through human interaction—and even from clickstream data—are incomplete, generally inaccessible to other applications that might benefit from the data, and difficult to analyze.

Data obtained from chatbot interactions can overcome these shortcomings. Financial institutions must make digital assistants part of their data management strategies – not just their sales and service strategies.