• Rajashekara V. Maiya, VP, Global Head-Business Consulting at Infosys Finacle


• 13.10.2020 08:12 pm

At the end of last year, Infosys published a 2020 trends forecast looking at Artificial Intelligence in banking, but much like everyone else, COVID-19 has forced us to reassess that. A significant change is that banks – and industry, in general – will assert their demands from Artificial Intelligence (AI) technology, rather than waiting for AI vendors to propose solutions. These demands will be of a structural nature, and more sophisticated to ensure relevant applications in future use cases.

Banks will expect so much more

If there’s one thing the pandemic has taught us, it is the importance of resilience; it is clear that only resilient businesses will last through the economic lockdown. So, this is likely to be banks’ first expectation from AI – that it helps them become stronger and more enduring in the face of tough odds.

AI must also provide re-traceability and be auditable itself. This will address some of the concerns around “rogue AI” – most important given that AI adoption will accelerate post-COVID.

The industry will also certainly expect AI deployments to be repeatable, across companies, countries and even continents. Since AI adoption is a continuous journey, banks will need the technology to be revisable from time to time, in response to new market and regulatory needs. The importance of having a system that is open to change was underscored in April 2020 when U.S. banks floundered in the deluge of stimulus checks issued by the Government as COVID relief; their systems were simply not ready to be scaled up to handle the surge. More problems arose for banks with the revision to Reg D removing the cap of six payments or cash withdrawals per month for savings and money market accounts. Banks, which never expected this longstanding rule to be amended, had hard-coded it into their legacy systems, and so struggled to implement the change quickly. Hence, all future applications and systems, including AI, must be flexible, configurable and parameterised to support change at speed and scale.

The next expectation from AI is reimagination – of products, processes, training & education, front office activity, services delivery and operations. The standard operating procedures of old will cease to be relevant in a post-COVID world and shall need to be rewritten. The same applies to products, which will need to be reimagined for end-to-end digital onboarding, contactless fulfillment, new interest regimes and so on.

The financial crisis of 2008 led to regulations to manage liquidity risk. Given the massive number of business failures and job losses in the coronavirus crisis, there is now a need to manage solvency risk as well. Banks believe that AI should help them manage the risks in business better, particularly the risk of insolvency, both in their own and their clients’ organisations.

While AI may take time to meet these expectations, it will in the short-term influence the following post-COVID shifts in the banking industry.

Consume less; produce more

It is quite clear that many of the branches that have closed for the pandemic will not open again: one report says this figure could be as high as 30 percent in Europe. With fewer branches and shrinking footfalls, banks will see a spike in transactions on digital channels. AI will have a huge role to play in maintaining the quality of experience (and advisory) on these channels.

Thanks to “no-touch” and social distancing, face-to-face interactions with relationship Managers (RM) will also dwindle, to be replaced by video calls with RMs, who may be assisted by chatbots armed with tools and insights promoting better, more confident, engagement. In this way, AI will help banks achieve increased communication in spite of less interaction.

AI will also help banks distil their unnecessarily long product lists into a smaller range of reimagined, risk-managed offerings. By making products more relevant to the post-COVID scenario, and simplifying associated processes, banks will end up doing the same amount of business, if not more, with fewer products. 

Forget selective; adopt comprehensive

All the solutions under the AI umbrella will come together to achieve comprehensive banking automation. Machine learning (ML), deep learning (DL), chatbots, robotic process automation (RPA) and advanced insights will deliver fully digital banking, including digital onboarding, digital transacting, digital delivery, and even digital closure (which is still manual and paper intensive).  And digitisation will not be limited to financial transactions alone; even routine queries that were previously handled in a branch or call centre will be attended to by a chatbot with complete context and traceability.

Broaden the use cases

Banks will witness broader, more sophisticated application of AI in common use cases, going forward.

• ML for enhanced security: As digital transactions and consequently, cyber risks increase, AI will be used to enhance banking security to new levels. (Think voice-based customer authentication using machine learning solutions that can detect voice variations with time of day.) In corporate banking, voice may be used to authorise bank guarantees, letters of credit etc.
• Deep learing to enhance fraud prevention: DL will support rapid risk scoring to ensure genuine transactions are not rejected. The solution will use transaction data of the past 5 to 10 years to respond in milliseconds.
• Enforced (Robotic Process) Automation: With manpower and budgetary resources drying up in the pandemic, banks will be forced to automate to stay afloat. In the near future, RPA will be deployed extensively across the front, middle and back office, and at all touch points.
• Hyper localization on the Internet of Things (IoT): By connecting with IoT devices, AI will enable banks to ramp up product sales locally; using information about preferences and location, it will present hyper-localised, hyper-personalised offers to customers.
• Bots on unassisted channels: Employing “regular” conversation and even urban slang, intelligent chatbots will simulate human-to-human interactions to ease the migration to unassisted banking channels.

Rise above the front-middle-back office

AI will soon rise from an operational level to a strategic one, where it will amplify strategic management, from strategy planning and formulation to implementation and decision making. This means that banks’ leadership and top management will also be more hands-on with the technology in post-pandemic times.  AI’s time is here and the pandemic has accelerated interest in its application to every area of the banking process – not just trading.  Banks know what they want from AI, moreso than other sectors, and it will be fascinating to see how they use it innovate and compete over the comings years.