Realizing the promise of AI in banking

This article was first published by Oliver Wyman here.


Artificial intelligence (AI), along with its subsets of machine learning, and advanced data analytics, holds the promise of revolutionizing the banking industry. Deployed at scale, AI can transform banking by enabling deeper customer relationships, sparking sales growth, and simplifying operations to bring costs down.

To date, however, most banks have only timidly ventured into the AI space with proof-of- concept use cases. Some of the players making best use of AI in banking are technology companies—fintechs, digital banks, and largescale telecoms. And consumers are taking note. A 2021 Oracle Global Retail Banking Consumer Trends report notes that 56% of respondents would likely switch to a banking proposition provided by Google or Apple.

Traditional banks are threatened by these new and often more agile entrants. But the largest financial institutions still enjoy an edge in customer data, marketing power, client trust, and distribution, and have an opportunity to leverage these strengths if they move quickly.

The power of AI

How can AI help banks regain their competitive edge? AI makes it possible to minimize manual intervention in core processes. AI further makes it possible to introduce hyper-personalized propositions and can transform key functions such as risk.

Our recent work with a leading bank shows the introduction of AI in selecting campaign leads has had remarkable results: 10% of all recent sales for personal and auto loans are linked to AI campaigns. Also, leveraging advanced data analytics to optimize pricing yielded up to a 5% increase in net income in personal loans.

Examples of the power of AI abound. One organization has improved detection of fraudulent transactions by 200%, by deploying an AI-enabled fraud-detection engine. At another, the use of machine-learning models to leverage social-media helped to quadruple digital cross-sell. Another institution has reduced client service and support costs by implementing an in-app chatbot that can service millions of clients. A European trailblazer achieved a perfect score for client experience once it embedded AI in its app—which made it possible for the institution to assess clients’ financial health and use that information to offer tailored plans.

Pillars to AI success

While many banks have indicated their intention to embrace AI, few have truly done so. Many remain too focused on day-to-day operations to invest sufficient time and effort on an AI transformation. Limited attention to AI spells limited success.

We have identified four pillars that trailblazers in AI transformation are following:

Deploy AI everywhere. AI can make a substantial improvement in all areas of the banking value chain, from the front office to the back. Start with applications that can deliver value quickly, as this will win support and hasten the adoption of AI in other areas.

Stand-out trailblazers with AI across all operations have the advantage of financial and strategic value creation. A holistic approach to implementing AI will allow banks to excel in client experience, support their business goals, empower decisionmakers with more and higher- quality information, and automate operations.

Work like a tech company. To compete with tech companies, banks need to mimic their agility, the linchpin of connection to customers. To truly succeed, banks will have to change the way they work, uprooting entrenched “silo” behavior and encouraging a multidisciplinary approach that allows for greater speed and efficiency.

Consider one large institution that customizes plans for clients. Having analyzed all available data, the bank will call on the business side to define the offer and the marketing team to develop digital support channels. Even risk and financial support will be required to play a role in approving the offer. This is a multistakeholder undertaking—vastly different from traditional models that see functions working in a sequential, siloed manner. Implementing AI without an agile approach is unlikely to be successful.

Transform the technology behind the data brain. A successful AI transformation feeds on high quality “raw material”—and good data is the feedstock for AI. Our analysis shows that trailblazers have implemented data-transformation programs that strengthen data governance, instill robust data management and data quality-assurance frameworks, and overhaul their technology core.

This does not mean all technology and data elements must be in place before a company can embark on the AI-transformation journey. To the contrary, our experience has shown that the project is likely to be more successful if the bank first concentrates on launching use cases that demonstrate the benefits, to elicit support from the broader organization.

Equal attention must be given to transitioning to the cloud, the most efficient channel for building the required computational power. A cloud transformation presents challenges but adds value beyond measure by exponentially increasing AI possibilities.

Build an “AI factory” to deploy AI at scale. Assembling a team of high-caliber individuals and formalizing a center of excellence to drive system-wide implementation is key.

In general, banks have not done a good job at this. Rather than filling their teams with data scientists and engineers, they appoint small, under-skilled teams to develop use cases that are often disconnected from the broader organization.

This significant stumbling block can only be overcome by creating an “AI factory” in which high-caliber individuals drive high-performing teams. Trailblazers’ data teams include specialized units with employees whose skills are regularly sought by the likes of Google and Facebook. One trailblazer based in Europe acquired a specialized big-data startup, the start of what today is a 200 member-strong AI unit that includes many PhDs.

Follow the trailblazers

Many attributes have contributed to the longevity of the traditional banking model. However, as consumers become accustomed to the superior value proposition made possible by AI, the value of the traditional model declines—rapidly. Banks cannot hope to remain relevant without harnessing the competitive edge made possible by AI.

Traditional banks should seek to emulate trailblazers—by embracing a holistic approach to AI that remodels the bank as a technology company in the banking industry. This is a truly disruptive exercise, but one that can be made easier by putting in place the building blocks—the “four pillars”—that that have propelled the trailblazers to their successful AI transformations.