Eight considerations for conversational AI in banking
By Pavithra R
Artificial Intelligence (AI) can be financially and technologically accessible to banks of all sizes, and a competitive advantage for banks looking to compete with fintech and large banks in delivering a superior customer experience. Yet, there are several best practices that banks should take into account to increase their chance of success in an AI implementation. According to Finn AI, these practices are:
- Start with a pre-trained banking system, not a general-purpose AI platform: Training, not software development, is the largest expense in creating an AI chatbot. A banking trained AI system knows that a “branch” means a bank building, not a tree part. And it knows “card” means a debit or credit card, not a part of a poker hand.
- Choose an industry-specific managed service for a significant reduction in initial and long term lifecycle cost: Developing from an open-source AI foundation puts the entire training burden for the life of the project on the bank, and the model needs to be trained on everything banking specific. An alternative method is to use a managed service; specifically, one built for banking. This offloads the perpetual training and takes advantage of banking knowledge already built into the model.
- Use AI to automate the contact center and deliver faster service for routine tasks: By u AI-powered bots to handle contact center volume, banks can achieve increased efficiency, offer more accessible service and scale quickly. According to Finn AI, automating the contact center with AI has the potential to offload 75% incoming requests.
- It is critical that a chatbot has a built-in understanding of a bank’s products: A highly functional banking chatbot will combine the concepts of pre-trained Machine Learning, with a pre-programmed hierarchical understanding of the specific product offering at the bank.
- Connect the chatbot to digital banking for maximum value: Transactional questions, such as transfers, balance inquiries and payments, will require both authentication and connectivity to the digital banking system, to access account data. To access digital banking data, the same set of APIs used by the mobile banking app can be utilized by the chatbot.
- Use AI to create a symbiotic relationship between bots and humans: Companies should design AI to take on certain repetitive tasks but which is ready to “handoff” more complex queries where a customer needs a human’s help. To accomplish this, banks need to establish handoff procedures that best meet the customer journey with the highest satisfaction rate.
- AI chatbots enable self-service, the desired option for most customers: Customer experience (CX) is moving to the top of the priority list for nearly all banks and credit unions. To give the customer their desired experience, a conversational bot option serves the customers who prefer natural language to a digital banking app when trying to self-serve.
- Chatbot position will play a key role in defining project success: To get to the front and center, banks should set up a schedule and target milestones for the chatbot’s ability to answer and contain requests, knowing that the topics discussed with an AI chatbot will be highly dependent on the page where the bot is found.
“Throughout the COVID-19 crisis, many organizations have doubled down on their investments in AI, likely because AI can deliver near-immediate cost and efficiency savings. Yet, while considerable effort has been made to make AI less of a black box and approachable to decision-makers, banks still want to better understand all the factors at play when making a product decision,” said Jake Tyler, CEO of Finn AI.
Also, read: Global FinTech Use Cases in Financial services
IBSi FinTech Journal
- Most trusted FinTech journal since 1991
- Digital monthly issue
- 60+ pages of research, analysis, interviews, opinions, and rankings
- Global coverage