17 Dec

Where Are ML & AI Being Applied Across Investment Banking?

Tom Jenkin

Financial services and specifically investment banking, has continually adapted to the changing socio-economic landscape and often leveraged technology to innovate how it drives financial markets across the globe.

In the world of high stakes investment banking, machine learning (ML) & artificial intelligence (AI) can be broadly deployed across the front, middle and back-office functions. As well as across various asset classes where decision-making processes can be enhanced with enriched and predictive analytics. The opportunities to improve risk management practices, create value for customers and enhance the consistency of making the correct decisions for the firm and its investors, are becoming increasingly unlocked by the day.

Investment banking is a somewhat incestuous industry. Where quid pro quo has become the norm. Ultimately ML and AI will become increasingly widespread as industry peers adopt and try to replicate the successes of their competition. Naturally, over time, this will evolve to a state where each firm will be reliant on the accuracy and performance of its ML & AI capabilities to remain ahead of its competitors.

As a ferociously competitive industry, investment banks can use intelligent tools and decision-making solutions to amplify each facet of the trade lifecycle. Whether this is through the creation of structured products, projecting value at risk, forecasting revenues on capital or enhancing accurate and speedy decision making.

We will go on to explore some of these areas in the next passage of this blog. Specifically;

- Algorithmic Trading & Quantitative Risk Management

- Anti Money Laundering (AML)

10 Ways AI is Reimagining the Financial Services Sector

Algorithmic Trading & Quantitative Risk Management

The Challenge

On a day by day basis, firms could be trading millions if not billions of dollars across various funds, asset classes and products. The frequency and volatility of these trades can vary across low latency and high volume demands. To trades that have fewer occurrences but require massive computational power so that a plethora of data sources across the bank’s technology estate, as well as trusted market data, news and media feeds, can be analysed in real-time.

Take those considerations across a single trade and dial them up across multiple regions, markets, institutions, asset classes and products. Very quickly you will be able to comprehend the complexity of the task that firms face in this space!

Furthermore, regulatory challenges such as MiFID II and FRTB have created additional compliance burdens for firms, as a result of the 2008 financial crisis. Calculating the potential return on investment and being able to project the risk that a trade could bring to the organisation's balance sheet is therefore imperative to secure new business, retain customer confidence and remain ahead of trading desks from other firms.

The Opportunity

In response to these challenges, firms are able to tap into ML & AI capabilities to support actionable intelligence that enables them to make more powerful trades, with sounder risk-based judgement. As an example, machine learning models can be deployed across structured and unstructured data sources to understand areas such as market and customer sentiment to determine potential risks associated with a specific trade that may not be immediately apparent to the human eye.

Algorithmic trading is increasingly becoming a dominant force across the global financial markets. As far back as 2017, JPMC reported that they had developed a first of its kind “robot” that had the artificial intelligence capabilities to perform buying and selling activities across its equities business, without moving market prices.

ML allows firms to closely monitor market conditions and understand events that can cause stock prices to go up or down. Deploying ML enables firms to analyse multiple internal and 3rd party data sources in parallel, with the intention of giving traders a distinct advantage over the market. This can bring other material benefits in the form of:

1) Increased accuracy and reduced chances of fat finger mistakes from traders.

2) Enable trades to be executed at the best possible prices, with a true understanding of risks.

3) Human errors are likely to be reduced substantially, with automated straight-through processing (STP) capabilities.

4) Enables the automatic and simultaneous checking of multiple market conditions.

Furthermore, as firms become increasingly accustomed to the fast pace of innovation that ML & AI can bring, further opportunities to push the limits in this space will no doubt evolve.

For example, by harnessing their use of data, firms should be able to further personalise the digital experience for their customers across their web and mobile-based investment platforms. A further example taken from JPMC can be found in the firm’s production of ~10,000 pieces of market research on a year by year basis. However, many customers were not able to locate this research and apply it as part of their investment strategies. By using ML & AI to analyse the portfolios of its customers, JPMC is now able to better match the research to customers through a recommendation and matching engine.

Anti Money Laundering

The Challenge

In recent years, anti-money laundering (AML) has been brought to the forefront for banks and other financial institutions for a number of reasons. Increasingly, there is a heightened focus on the environmental, social and governance (ESG) practices of financial institutions across the world. Firms are now being increasingly encouraged by regulators, their customers and society as a whole, to be more transparent about with whom and how they do business. Many firms are intent on building lasting brands that are associated with financial prosperity, great service and trust.

Money laundering practices can often be traced back to organised crime, terrorism and other illicit practices that ultimately have a detrimental impact on society. Whilst criminals and their approaches to executing money laundering practices are evolving in an accelerated manner. With new attack vectors evolving every day, the detection of money laundering is a big concern for banks across the globe. The reputational damage that can be brought onto a firm that is found to have not applied the correct AML policies and controls, is potentially catastrophic.

Traditionally, the processes to impose and enforce AML detection have relied too heavily on manual, human-intensive effort. Making these policies and controls challenging to scale and expensive to operate on a global stage. To reinforce the magnitude of this problem, it is estimated that financial institutions spend upwards of $214 billion on fighting money laundering each year.

Furthermore, compliance functions are under pressure to have multiple touchpoints with customers to validate who they say they are, when onboarding them in the bank’s ecosystem of services. Whilst it is even more challenging to demonstrate regular reviews of customers once they are unboarded and perform pattern detections for any spurious financial activities across their portfolio of products.

The Opportunity

In recent years there has been an explosion of FinTech and RegTech solutions that address a wide variety of Know Your Customer (KYC) and AML issues for financial institutions. Data, analytics, ML & AI are central to powering these solutions and they are becoming central to how firms respond to AML legislation that has proven to be an albatross around their necks for some time.

As part of many modern onboarding processes in the retail banking estate, whenever a customer wishes to open a bank account they will undergo a series of identity & verification (ID&V) checks to ensure completeness and accuracy of the customers. This may include facial and voice recognition technologies being cross-referenced against an approved form of identification (passport, drivers license).

Similar processes are executed in the investment banking domain where illicit funds can be freely moved through the financial markets if the requisite ID&V practices are not executed. This may include the geographical location of the customer, whether they are politically exposed, have ties to organised crime or if they are under any form of embargo or sanction. Automating these checks against trusted internal and 3rd party sources can quickly identify low and high-risk customers and trigger additional workflows that may require more meticulous human analysis.

In addition, trade or transaction monitoring can be used to identify suspicious trading behaviours that may indicate malicious intent to deceive the firm, the wider market and law enforcement agencies. For instance; if a customer is accepting inbound payments from multiple sources on their e-trading account, then performing a series of trades and pushing any profit or remaining funds to an alternative account in a high-risk jurisdiction then there could be money laundering practices at play.

Over time, as firms are able to use ML & AI capabilities to quickly identify, filter and establish known patterns that indicate money laundering practices, they have every opportunity to further optimise their risk management and oversight practices. Analytics and business intelligence dashboards can enrich a firm’s understanding as to where they carry the most risk across their institution and various entities.

Ultimately, this allows them to establish proactive and predictive insights around; suspicious activity reports (SAR’s), politically exposed people (PEP’s), high-risk clients, incomplete customer data-sets, active regulatory complaints or any outstanding AML alerts that have not been acted upon.

In Closing

Over the course of this blog, we have discussed the investment banking sector and highlighted the potential for the deployment of ML & AI. Having discussed algorithmic trading, quantitative risk management and AML we can conclude that:

1) ML & AI can tackle AML management overheads and potentially reduce costs & expenditure in this space.

2) ML & AI can be applied to increase a firm's ability to demonstrate transparency in-line with ESG targets and regulatory pressures.

3) ML & AI can be applied to better manage risk management practices and STP practices.

4) The opportunities to better understand risk vectors and market sentiment, enable firms to execute better trading positions. As a result, ML & AI should be central to their technology strategy.

In our next blog, we will delve into the opportunities that the personal and commercial insurance sectors have in respect of deploying ML & AI.

Interested in seeing our latest blogs as soon as they get released? Sign up for our newsletter using the form below, and also follow us on LinkedIn.

Latest Stories

See More
This website uses cookies to maximize your experience and help us to understand how we can improve it. By clicking 'Accept', you consent to the use of these cookies. If you would like to manage your cookie settings, you can control this in your internet browser. Find out more in our Privacy Policy