The concept of a cognitive corporate bank would have been considered moonshot a decade ago. Over the past three years, the use of advanced artificial intelligence (AI) in corporate banking has been accelerating. This progress enables us to envision what a cognitive corporate bank could look like, that is, one that optimally applies advanced AI as part of a broader digitization strategy involving other automation technologies.
A cognitive bank is not one filled with robot-bankers but rather is one that effectively applies AI when it is the optimal solution and enables bankers to focus less on administration and more on adding value. Over the next decade, AI will not only dramatically change the nature of bankers’ work as well as that of their commercial customers but also will reshape competitive differentiators.
Why now? After decades of work in academia and labs, significant AI breakthroughs are being achieved and commercial applications growing. A clear signal is that AI references in patent filings started skyrocketing in 2013. The rocket boosters have been cheaper, faster computing power and data storage and newly accessible data sources coupled with the commercial drive to advance machine learning and deep learning. Tech and digital giants, as well as banks, have taken on the moonshot challenge with some already making landings.
“The rocket boosters have been cheaper, faster computing power and data storage and newly accessible data sources coupled with the commercial drive to advance machine learning and deep learning.”
Advanced AI will have a profound impact on how humans do their work and engage with machines. It will transform relationship manager and customer engagement. As consumers, we are becoming increasingly comfortable talking to machines. This comfort level will gradually flow over to the business world as it did with prior innovations, such as mobile banking. A treasurer will ask a virtual assistant for FX exposures, risk assessment, and hedging options. A procurement head will ask a virtual assistant to generate a report on potential discounts available and support a negotiation process.
An accounts receivable manager will ask for a chart showing days sales outstanding for a specific customer vertical. A moon landing has already occurred with the first step by Sage’s accounting virtual assistant, Pegg. Bank relationship managers will reach out to clients based on next best action recommendations generated by a machine learning model.
AI implementation will be gradual, beginning with test missions before advancing to more ambitious missions and ultimately the moon, that is, the cognitive corporate bank. Similar to advancements in online banking, virtual assistant functionality is starting with standard reporting and answering basic questions. It will progress to providing insights and alerts regarding basic financials and enabling payments initiation.
“AI implementation will be gradual, beginning with test missions before advancing to more ambitious missions and ultimately the moon, that is, the cognitive corporate bank.”
Gradually, virtual assistants will be able to answer more sophisticated questions. For example, a six-month cash flow forecast based on the bank’s analysis of my accounts receivable and payable and its external macroeconomic forecasts. The “putting man on the moon” moment will be when a virtual assistant makes bespoke recommendations for responding to an alert.
For relationship managers, advanced AI will be increasingly embraced as a partner, increasing their productivity while making their jobs more interesting. They will be able to be more proactive in customer service when provided “next best action” recommendations based on analysis of a customer’s transaction behavior, customer communications, and external information and news.
“For relationship managers, advanced AI will be increasingly embraced as a partner, increasing their productivity while making their jobs more interesting.”
Machine learning and natural language understanding are enabling sentiment analysis of the customer as well as of external news sources. For example, it could detect customer dissatisfaction during his last call to the service center, which could trigger an outreach by the relationship manager. Or it could perceive a negative outlook for a customer’s business made by an analyst, which could trigger a review of the customer’s loans outstanding.
Relationship managers will share their desktops with their own virtual assistant, enabling them to delegate basics tasks. The concept of humans and machines working side by side has already been realized in compliance and fraud operations in which humans can review the work done by AI and focus on the cases that AI cannot resolve and through their actions train the machine.
Moonshot will become a reality only if critical prerequisites are in place. Successful adoption of AI rests on much more than the technology. Banks which want to break barriers must have in place a transformation-driven leadership which embraces new paradigms based on assumptions about the future rather than the facts of the past, a collaborative alchemy across the business-side, data science, IT, and compliance, an enterprise-wide AI initiatives team which ensures synergy and transparency, modern data infrastructure, and strong model governance and explainable AI to ensure ethical behavior, privacy, and robust permissions.
“Successful adoption of AI rests on much more than the technology.”
Implementing AI requires the vision, patience, perseverance, and collaboration of an Apollo mission team. Excelling in AI requires the fortitude and confidence of an astronaut. Banks that succeed in building the cognitive corporate bank will have braved going beyond the boundaries of traditional processes and embracing new paradigms, such as, using a cloud-hosted environment and letting machines do some of the talking.
As bankers explore the potential of AI, they should keep in mind what the famous pilot Chuck Yeager once said: “You don’t concentrate on risks. You concentrate on results.”