Financial institutions invested more in generative artificial intelligence (GAI) in 2023 than any other industry. In the final quarter of the year alone, NatWest and Commerzbank announced partnerships with IBM and Microsoft, respectively, to integrate the technology into their virtual assistants. In addition, JPMorgan Chase has filed trademark paperwork for a ChatGPT-like product tailored for use in financial services.
And there’s so much more to come. As GAI has matured from anticipation to deployment, in the process growing into a $166 billion market, many FIs are in full exploratory mode to find ways the technology can help reduce the cost of customer acquisition, servicing, compliance and fraud detection, while also boosting customer lifetime value (CLV).
In 2024, GAI is likely to grow unabated, promising to improve customer adoption, expand cross-functional personalization and automation, enhance immersive experiences and bring depth to customer relationships.
The chatbot was a key differentiator in 2023. According to Curinos’ Digital Banking Analyzer, nearly three-quarters of national banks and fintechs offered an in-app virtual assistant, compared to half of super-regional banks and only 21% of credit unions.
Within the main interface of the app or desktop, a basic chatbot provides links to key features, thereby keeping users within the chatbot experience and handing them off to a live agent only after triage. More advanced integrations complete requests such as displaying balances or recent transactions, and those at the forefront complete complex tasks, such as bill payments, Zelle transactions and card management.
Several chatbots deliver preconfigured responses using natural language processing (NLP), and GAI is enhancing the experience by making the responses personal. In 2024, we expect to see more banks using these conversational interactions to cue tools that move users closer to their financial goals, which in turn will increase CLV.
For Marketers, GAI
In marketing, artificial intelligence is already supercharging the test-and-learn process and better enabling predictive personalization. Its reinforcement-learning algorithms allow marketers to bypass ineffective A/B testing and evolve past basic propensity models that require regular refreshing.
By the end of 2023, marketers had used AI to continuously optimize customer journeys by message or channel type, tone, sequence and cadence, and predefined outcomes. One leading AI platform delivers personalized treatments that drive response improvements and has found that creative optimization alone can drive a 60% to 120% lift to existing decision-engine performance. Customers receiving optimized marketing are 1.5 times more likely to open a new product and twice as likely to engage with digital features. Accounts opened by customers receiving optimized marketing are more likely to still be open three and six months later. The platform was automatically optimizing more than 107 billion decisions per week by year-end, and the volume is expected to grow to half a trillion decisions per week by the end of 2024.
Financial institutions using these types of systems are currently fulfilling content needs, and as capabilities grow, some will likely look to GAI as well. The technology’s ability to generate easily customizable and human-like creatives stands to vastly improve productivity around content production. By doing so, banks have the potential to integrate proactive and conversational insight delivery, and expand performance marketing creative development and optimization, which ultimately lowers cost per acquisition and further deepens relationships.
Staying Ahead Of The Bad Guys
Like so many emerging technologies, generative AI will be used for both good and evil.
Many retail banks, for example, have streamlined onboarding and account access with biometric login functionalities and multifactor authentication systems. But GAI has already allowed attackers to breach the systems with convincing impersonations by phone and email and by bypassing camera feeds through synthetic imagery and video recordings and deepfakes. As GAI comes of age, facial, voice and fingerprint deception will be increasingly available to hackers, putting banks at increased risk.
In 2024, banks will be turning to predictive analytics to ward off such fraudulent activity. One emerging example is in the payments space, where large transaction models (LTMs) have been built and trained using GAI algorithms to identify hidden transaction patterns – such as time sequencing – that current methods often can’t detect. Using these algorithms and huge volumes of data, machines will be able to efficiently identify the relationships between varying customer transactions and assess their level of risk.
The Need For Accuracy And Trust
Perhaps the biggest impediment to further adoption of GAI in 2024 is the concern, by both banks and regulators, of “hallucinations” – inaccurate information delivered through chatbots or content generators.
The Consumer Financial Protection Bureau noted in 2023 that untested or poorly managed GAI distribution, particularly in chatbots, could lead to noncompliance with federal consumer financial protection laws while harming an institution’s trust among customers. Those who are keen to get a competitive edge, stay on the right side of the regulators and fulfill customer requirements with GAI and its ability to create content are doing so with the guidance of a human hand.
While predictive AI is now being used to fully optimize customer experiences and deliverables, many more GAI integrations are happening away from the customer. Along with Know Your Customer and security, one major U.S. bank recently announced that it would be integrating the technology into interpersonal sales and advisory services, whereby advisors are provided with investment options in real time.
However, customers will begin to engage more directly with GAI, and it will happen soon. In the near term, we will see the production of visuals that help inform each customer’s decisions around their financial health and direct them through journeys in-channel. In the medium term, GAI will be employed to perform an increasing list of tasks on behalf of the user, such as raising queries and filing documentation. Longer term, advocates foresee GAI-sourced content playing more of a direct hand in sales execution, which is already taking off in less regulated industries like retail.
As we look deeper into 2024, we see financial institutions accelerating their investigation and development of GAI so they can effectively put it to its best use for their customers and clients, and for themselves as well.