As per ACFE’s last Report to the Nations, fraud is a staggering $ 4 trillion global problem – a size and scale astronomical enough for every bank to seriously rethink their entire fraud risk management approach. Let’s take a look at the top 5 implementable best practices that make for the foundation of a robust banking enterprise anti-fraud framework.

5. Realtime fraud prevention across channels
We are four fifths the distance away from the 22nd century as of writing this, even closer when you are reading this. While we have come to realize that multi-channel banking has come to stay, we need to realize that the future will not just be multi-channel but also an unknown number of and unknown combinations of multiple channels.

Today we think in terms of a finite number of channels – ATMs, POS terminals, mobile banking, internet banking and of course the good old branch banking.

However with open protocols becoming the norm, some actively implemented and growing exponentially, like India’s UPI, where vendors can create their own apps to interact with multiple banks, to the Open Banking API platform, in the very near future, you would have a channel for banking that was never even in existence, let alone imagined, with its own balance (or lack of it) between convenience and security.

Guarding fraud in real-time across channels in such a scenario requires guarding the sanctum sanctorum rather than all the practically infinite pathways that lead to it. Guarding those channels individually will just not be scalable nor worth the expense.

4. Sure, start with heuristics, but refine them via automated classification
Your bank has the domain knowledge – heuristics – of over a cumulative years of fraud detection and prevention expertise lying with your seasoned employees. This is a good place to start, when configuring your fraud prevention system to get easy wins in the beginning – a low hanging fruit.

However, automated tools, that have the advantage of being able to process terabytes, petabytes and exabytes of data to arrive at new insights to classify all the actors in any transaction, into possible fraud victims and possible fraudulent actors – must be taken advantage of. These tools are getting commoditized by the day and most fraud prevention systems offer you to start with heuristics and then move on to automated classification.

This is necessary for managing the sheer volumes of transactions. Automated classification acts as a powerful magnet that helps discover the ‘needle in the haystack’ faster.

3. Catch them, then remember them
So you prevented a fraud, then showed management the dollar savings and earned praise. Great. But did you know that you can also showcase further prevention by remembering the entities (humans, machines, locations, time of event, frequency and velocity of events) and applying those patterns to prevent further potential frauds?

Any fraud prevention system that doesn’t take existing fraud parameters into account to improve the prevention algorithm is just doing 10% of the job.

2. Profile victims to prevent similar preying patterns
What goes for fraudsters, goes for victims too. Do you find a particular fraud prevalent with, say, a family of 1 or 2 with ages over 65, residing in smaller cities and/or living on monthly pensions? Or with young people in their late teens / early 20s with heavy social media usage, receiving monthly allowances from parents?

If they could scam one of them, you can be pretty sure that the trap has been laid in other similar places. A group fraud detection system can uncover patterns that can help identify potential victims with a similar background. Profile your victims so that you safeguard your customers while reducing your risk exposure.

1. Maintain the balance between policing and good service
In the age of instant likes and viral everything, no one wants to wait for more than a minute for any digital banking service. All your efforts to provide a secure transaction environment for your customers is not of much use to business if customers end up becoming ex-customers.

So, real-time fraud prevention that doesn’t increase your service latency by more than 100 milliseconds (yes, consumers can sense that half second delay and it can lead to a 20% drop in usage) – is ideal.

Ask for benchmark results from your fraud detection vendors and get the right references before you arrive at a purchase or subscription decision. You want to prevent fraud, but you don’t want to lose customers either.

A holistic approach to targeting fraud in real-time with a blend of heuristic and automated learning is necessary to achieve optimum ROI from your fraud prevention investments.

Also, the banking sector is scaling at a rapid pace to service existing and new customers (including those who may be digital banking newbies or may not be very tech-savvy). This, when combined with accelerated operational automation, diminishing human contact and instances of newer ‘innovative’ fraud schemes, we see there is a clear and present need to re-imagine how financial crime risk can be managed, without impacting customer experience, across the banking enterprise.

To begin with, applying these best practices will be a sure step forward towards laying the foundation of an intelligent banking enterprise anti-fraud framework.

Business intelligence (BI) solutions are not new to banking. With a blend of tools, techniques, and technology, BI processes information with accuracy, which otherwise would have been performed by humans. BI solutions have been enabling banks with capabilities to generate more revenue, reduce costs, mitigate risks, and more.

Driven by changing business models, disruptive technologies, compliance pressures, and new competition from fintechs, banking and financial services is undergoing a radical shift including the way they are using BI.

The adoption of BI applications in banking has helped improve efficiencies while lowering operational costs. Despite BI having come a long way, the challenges of data silos, security concerns and compliance have remained.

Sustainable success in business demands real-time insights, agility, exponentially growing customer relationships, and continuous innovation. Banks today need to be able to leverage the agility and scalability of contemporary technology innovations to stay ahead of the game.

BI and AI are distinct but complementary. The ‘intelligence’ in AI refers to computer intelligence, while in BI it refers to the more intelligent business decision-making that data analysis and visualization can provide. BI helps banks put the proverbial ‘method in the madness’ to the massive amounts of data they collect. But banks need more than smart visualizations and dashboards in this day and age.

Interestingly, AI and BI have significant, and in certain cases, complementary enterprise applications.

AI can enable BI tools to deliver clear, useful insights from the data they analyze. An AI-powered BI system can clarify the importance of each datapoint at a granular level, and help human operators understand how that data can translate into real business decisions. With the convergence of AI and BI, banks can synthesize vast quantities of data into synchronized action plans.

AI Transforming Decision-Making
AI-based BI tools are simpler to use, deliver more useful insights, and make business users more productive, saving substantial time and money. More than just provide superficial insights, AI-based solutions recommend ways to fix issues, run simulations to optimize processes, create new performance targets based on forecasts, and take action automatically

AI is making a substantial difference in analytics by democratizing data and improving adoption. With their extraordinary ability to adapt and learn, AI-powered solutions deliver better real-time insights and feedback data.

Core AI applications such as predictive analytics and ML have opened the doors to a whole new generation of applications. Since AI can analyze massive quantities of data and deliver instant recommendations, a fundamental shift is already occurring, and its impact is evident across industries especially banking and financial services.

Insights are now much more accessible and comprehensible to the average user, and with such an upgrade, they can assist business in reconsidering their strategies.

Banks and AI – where are we today?
Today most banks and credit unions are only beginning to use AI. Over 30% of financial services use AI technologies such as predictive analytics, recommendation engines, voice recognition and response. The most ubiquitous form of AI is chatbots that are actively used in the front office. The top areas using AI remain cards and payments.

A recent analysis of very select ivy league banks in the US revealed that AI is being used in complex, repetitive processes and data analysis. A leading global bank recently launched a contract intelligence platform for analyzing legal documents and extracting important data points and clauses.

Their ML technology enables analyzing thousands of commercial agreements in a few seconds. This means a dramatic reduction in the time spent on manual back-end processes.

Another leading US bank adopted AI for its intelligent virtual assistant, which uses predictive analytics and cognitive messaging to provide financial guidance to their 45 million+ customers.

Yet another business-critical application of AI in banking is fraud identification and prevention. With the average loss to fraud attacks at 1.5% of their annual revenue, one of the most important application of AI in financial services is risk management,

A leading global US bank partners with tech companies to improve its services and stay in the forefront. One of its strategic investments is using ML to evaluate “big data” and potentially fraudulent activities in all avenues of commerce including online and offline banking.

AI Reshaping Risk Management
AI has turned out to be a game changer for fraud risk management in finance as it provides banks and credit unions with the necessary tools and solutions to identify potential risks and fraud.

The financial crisis of the previous decade gave financial services firms a lot of problems with credit challenged consumers.

Before the digital revolution in the financial services industry, customer intelligence was based on some relatively simple heuristics, the customer value data was gained via focus groups and surveys of consumer behavior – results of which didn’t necessarily convey reality.

Today new technologies provide banks access to vast amounts of data about consumers’ behavior and needs.

Another great source of problems for banks is online lending technology and the emergence of alternative lenders. Non-traditional lenders use technology-based algorithms and software integrations to assess credit profiles of customers and are also leveraging alternative data such as social media photos and check-ins, GPS data, e-commerce and online purchases, mobile data, and bill payments.

Risk management in banks must use cognitive technologies to gain competitive advantage and use risk to power their organizations’ performance.

AI in Financial Crime Risk Management 
Given the development of digital technologies and shrinking costs of data storage, AI is now becoming an integral part of business processes. ML allows managing and analyzing unstructured data, saving time and money of financial services companies.

AI in banking risk management lowers operational, regulatory and compliance costs and provide reliable credit scorings for credit decision makers. Risk assessment AI can provide a fast and accurate risk assessment, using both financial and non-financial data to factor in the character and capacity of a customer.

AI-powered risk management solutions are also used for model risk management (back-testing and model validation) and stress testing, a key requirement by European and US regulators.

AI-based Risk Management in Banking – What Are The Challenges?
Besides the technical challenges of developing AI apps for banking, e.g. building correct and relevant algorithms, there are also challenges related to regulatory field and data access rights.

The fintech ecosystem is governed by strict compliance to regulations related to data. Data breaches can be expensive and new legislations, e.g. GDPR in EU, provisions for serious liability for companies dealing with personal data. Also, there are several regulations governing financial institutions in the US and Europe, e.g. FINRA, MiFID and EMIR.

The ability of ML models to analyze large amounts of data – both structured and unstructured – can improve analytical capabilities in risk management and compliance, allowing risk managers in financial institutions to identify risks in an effective and timely manner, make more informed decisions, and make banking less risky.

AI and risk management dovetail well when banks need to manage unstructured data. Risk managers must focus on analytics and stopping losses in a proactive manner based on AI findings, instead of spending time managing the risks inherent in the operational processes.

Moving from ‘What happened?’ to ‘What’ll happen next?
With unimaginable volumes of data generated each day via millions of transactions, banking is quite easily one of the most data-intensive sectors. The volume of data generated will only swell further, also due to the proliferation of newer technologies.

Banking in the BI age used insights to understand customer behavior, make financial forecasts, analyze operational performance efficiencies, and create reports for regulatory compliance. BI has now begun adopting features and capabilities that blend machine learning with conventional BI offerings.

With open banking and digital transformation having completely changed the game for financial services, it is imperative to monetize value from a continuously changing environment, and do it real fast.

The future of BI is beginning to be powered by AI. There are numerous applications of AI in the financial services ecosystem that help identify patterns and connections in real-time that humans cannot, thus opening the doors to newer possibilities to gain better customer intel, boost growth, and combat fraud.

Regardless of certain valid differences, AI-led BI is a force multiplier. Rather than view AI and BI as disparate technologies, banks must invest in realizing the combined potential that will only help them grow to newer heights.

References

• AI: The new BI
• Banks utilizing AI-driven BI
• Big Data in Financial Services – From data to insights
• Using AI in financial fraud detection can help banks save unnecessary losses
• ML & AI the future of BI?