Why Generative AI is the Bank’s Solution to the Big Tech Challenge

Don’t let the Chat GPT hype put you off: large language models (LLMs) and generative artificial intelligence (AI) are the secret weapons banks need to meet Big Tech’s threat.

That’s the view of fintech entrepreneurs Leon Gauhman, co-founder and CPO/CSO at digital product consultancy Elsewhich boasts customers included Spotify, Google, Microsoft and MasterCard.

With neobanks apps surpassing those of legacy banks and Apple and Google leading the race for digital wallet dominance, Gauham argues that embracing LLM will allow established banks to tip the scales in their favour.

If 2022’s overhyped topic of tech conversation was the metaverse, this year’s equivalent is OpenAI’s ChatGPT. At the height of the hysteria, it seemed possible that the powerful chatbot would surpass the roles of academics, journalists and even lawyers.

But the more we discover about the limitations of ChatGPT, it’s obvious that there’s still some way to iron out before these technologies can deliver on their revolutionary hype.

Incumbents and newcomers should still pay close attention because the key from the development of ChatGPT and similar generative AI tools lies within the capabilities of large language models (LLM) – the core technology that powers them. Microsoft and Google have both just launched LLM-powered workplace collaboration tools with Microsoft also investing $10 billion in OpenAI, the company behind ChatGPT.

The speed with which both companies are integrating LLMs into their businesses underscores the importance that big tech—an emerging competitor to incumbents and upstart banks—attributes to this technology.

With Apple and Google moving forward in the race for digital wallet dominance, new developments in AI-powered deep learning could allow incumbents and new banks to tip the scales in their favor, in the following four areas:

1) Customers can have their own AI personal banker

large language models play into one of the bank’s biggest advantages: its proprietary datasets. Using large amounts of customer data and insights, banks can implement deep learning and natural language processing tools to create their own valuable IP; rather than trying to replicate its rivals with piggyback products.

For example, a well-trained version of ChatGPT allows for the creation of a personal AI banker that provides customers with real-time recommendations perfectly tailored to their individual circumstances and needs. At a stroke, this could reinvent the bank’s reputation for online user experiences, and create space for genuinely innovative, responsive products that are in line with users’ expectations.

2) AI can increase banking productivity

Great language models aren’t just about customer experience: the technology can also help employees by augmenting the resources they have access to. For example, Microsoft claims that its Copilot tool will help Office users create presentations and prepare for meetings by providing relevant updates. Meanwhile, Google describes its AI tool as a “collaborator” that can suggest, summarize and provide insights.

More generally, these technologies can potentially revolutionize labour-intensive workflows around processes such as KYC, compliance and AML. With these core businesses accounting for 15-20% of banks’ budgets, the implications for increased productivity, freed up time and budget are enormous.

3. Big language models are the keys to digital transformation in high gear

LLMs have the potential to energize the entire financial services stack in support of digital banking, a reality that legacy players have struggled with to date. With new vendors set to build large open source models, banks will have a direct channel to train LLMs using their data. These developments will in turn allow banks to add a generative AI layer across a broad subsection of digital processes, from product design to mobile banking and cyber security to employee onboarding.

For example, Swedbank already uses generative resistance networks to identify fraudulent transactions – drawing on synthetic modeling to understand and predict under-the-radar irregularities. In the race for digital banking transformation, this capacity becomes a superpower, enabling banks to leapfrog more nimble competitors.

4) Large language models herald a new era of proprietary modeling

LLMs allow banks to take proprietary data, extract valuable insights from it and use resulting action points to develop new personalized banking/wealth management strategies or a conversational interface.

For example, LLMs can analyze ten years of bank data around mortgage defaults and use the findings to create a new, ever-learning underwriting framework to make better lending decisions.

Last thought

In these times of hyper-innovation, banks are often criticized for being too cautious and slow to adapt. When it comes to LLMs, a cautious approach is best. As we have seen Bing ChatGPT-powered search and Google’s AI chatbot Bard, AI models tend to “hallucinate”, generating biases or inaccuracies. An AI chatbot that hallucinates a wrong answer about NASA’s James Webb Space Telescope is relatively benign.

By contrast, a banking brand’s AI chatbot hallucinating inaccurate financial advice would be seriously bad news. This potential risk makes rigorous human input and oversight in data training stages a necessity, not an option.

But caution is no excuse for inaction. If banks want to future-proof their businesses against the threat of big tech, they need to move forward in understanding, experimenting with and training LLMs. The biggest risk for banks, evident with other recent technology pivots, is to do nothing. Now is the time to regain ground, using the LLM as a transformation agent to make customer centricity a core business.