IBM RegTech Innovations

The top five challenges and opportunities for AI in RegTech

Share this post:

In the span of a just few years, artificial intelligence (AI) has gone from a niche, relatively abstract concept, to entwining itself in multiple aspects of our daily lives. AI has found applications such as how we avoid traffic while navigating the physical world, to how we search for products online, to how we interact with virtual assistants in our homes and on our mobile phones. Based on the findings of a recent survey, it seems that risk professionals across the RegTech landscape are also awakening to the promise of AI, helping them obtain a better picture of the benefits within their organizations and assess where they themselves sit in its broader evolution.

The growing use of AI in financial services

Within the financial services industry, early adopters have already started to integrate these cognitive technologies into higher value customer service and account management applications, such as chat bots and virtual bankers. However, until recently, these advancements had been only marginally utilized within the risk and compliance groups. Whether due to lack of specific use cases, or limited visibility into the challenges analysts, compliance officers and risk managers face. But financial institutions are awakening to the potential impact these technologies encompassing AI can make – and regulators are on board as well.

Gauging AI by asking the experts

A few months ago, IBM conducted a survey of risk and compliance professionals to better understand which aspects of AI they were utilizing and in which applications. Out of that report came five major themes, both challenges and opportunities, in the use and adoption of AI in risk and compliance:

     1. Challenge: Lack of skills and data

A pervasive theme throughout the report, the limited number of staff that can use AI effectively and the lack of usable data will both slow the adoption and impact of AI. The primary groups using AI within financial institutions are focused on research and strategy or for very niche applications. These two factors are interconnected, as the limited data will hinder the number of use cases that can be explored with AI and the small number of professionals with the skillset to utilize AI will constrain the wider application of cognitive techniques.

     2. Opportunity: More usable AI is coming

As AI works its way into more applications beyond the current, targeted use cases, the benefits will likewise have a wider effect on the organization and for the risk and compliance groups as a whole. For example, rather than automating one aspect of an investigation or workflow process using natural language processing (NLP), AI techniques like NLP will be applied throughout the entire workflow, significantly improving the speed and effectiveness of analysts. This “packaging” of AI capabilities will diminish the need for the wider risk and compliance group to have the advanced skills needed to use AI, though those skills will continue to be in high demand.

     3. Challenge: and adoption initially

Of the topic areas covered, the primary drivers were time and cost savings as well as achieving compliance. In terms of the types of AI tools in use, more than 30% were utilizing natural language processing and machine learning either extensively or as a core component. More than 50% were not using AI techniques as segmentation, evolutionary algorithms and graph analytics, though most respondents were aware of these tools. While roughly two-thirds of respondents attested to using AI in their organizations, it seems both the applications and techniques are currently limited.

     4. Opportunity: Professionals recognize the wider value

One of the reasons for limited use of AI among certain financial institutions and professionals might not be that they don’t believe in the value of technology; on the contrary, these groups may have already invested in technology and been doing complex statistical analysis, which is why the need to shift to AI-based platforms is less urgent. For example, financial risk professionals tend to see AI as another tool rather than a transformative technology. In addition, these respondents were the only ones not to list inefficient skillset as their main challenge.

Likewise, GRC and financial crime risk management groups recognized the potential application for AI across a number of areas, starting with those that can gain efficiency and progressing to greater compliance effectiveness and detection accuracy. As the adoption of AI continues and professionals become more familiar with AI tools, the use cases will grow in complexity and further improve the value they bring to financial institutions.

     5. Opportunity: AI can simplify transparency and explainability

Regulators, auditors and boards are demanding more clarity from financial institutions, especially as AI becomes part of the decision-making process. While this may seem like a challenge or barrier to AI adoption, cognitive technologies if implemented correctly, can provide more consistency and transparency into the decision making process than analysts, especially in terms of subjective analysis like enhanced due diligence reviews. They can also automate the collection of the supporting evidence that led to the final decision, making the audit process less time consuming. In addition, as familiarity with AI grows, the collective understanding across the risk and compliance function will likewise grow.

The need for a collective approach to AI in RegTech

One last takeaway not in the survey is the amount of overlap across risk and compliance teams. Even though many of groups focused on governance, risk and compliance; financial risk management and financial crime risk management may have common goals and even use common tools, there has been little cross-pollination or collaboration, especially when it comes to the application of AI techniques or use cases.

In the risk and compliance space, it can be difficult to extract oneself from the pinpoint focus on one’s specific responsibility. But I would argue that taking a wider view of the space can provide not only a greater perspective on areas of overlap, but also interesting developments that can be leveraged to improve the way you manage market and credit risk, enable strong model governance or triage AML transaction monitoring alerts.

AI may be the latest in those developments, but protecting financial institutions requires a look beyond a single tool, focus area or technique to better understand where the collective use of regulatory technology can have an impact. Read the complete results of “AI in RegTech: a quiet upheaval” survey here.

Content Director, WFS Financial Crimes & Conduct Risk

More IBM RegTech Innovations stories

Payments modernization: Addressing threats and assessing new opportunities

Payments modernization initiatives around the globe have long been a topic of huge interest and are at different stages of implementation in different countries. Time and again, it has proven to be a catalyst for transformation that has fueled innovation. Canada is now completely into this modernization party, driven by the modernization agenda and timeline […]

Continue reading

Stop waiting to apply AI for financial crimes

Challenges from financial crime incidents and penalties grow Criminal enterprises are becoming increasingly sophisticated at the same time as new payment methods are emerging (peer to peer and faster/immediate payments) and transaction volumes are exploding. And financial institutions across the globe are feeling the pain of high-profile financial crime incidents in the form of significant […]

Continue reading

A new era of technology enabled financial risk management: Advanced analytics and aggregation

In this series of blogs, we focus on four high-growth transformative technologies with emerging risk applications which can help banks and financial institutions complete and protect the enterprise. Today we discuss how advanced analytics and aggregation software can address limitations in computational power and granularity required to meet evolving regulatory demands.The other three emerging technologies […]

Continue reading