Skip to main content

Combat financial crime with AI and advanced technology from Microsoft

Financial services organizations have long recognized technology as a transformative force in their business models. Now they’re at the cusp of taking advantage of new advances in AI and data science to seriously combat some of the most pernicious criminal activities around the world.

With Microsoft Cloud for Financial Services, our customers are managing financial services data at scale and building solutions that improve customer experiences and operational efficiencies. With the advent of generative AI capabilities in Azure OpenAI Service, businesses can now unlock new value from their data not only to drive better customer outcomes but also to improve their protection against various kinds of financial crime—including fraud, electronic crime, and money laundering.

The financial costs and scale of these crimes are staggering. Worldwide, the estimated total of laundered money in a year is at least two percent of global gross domestic product, or USD800 billion.1 For financial services organizations, the cost of financial crime compliance reached USD213.9 billion in 20212—USD56.7 billion in Canada and the United States alone in 2022,3 a 13.6 percent increase from 2021.

Until recently, financial services organizations have felt hamstrung in their ability to combat the worst forms of criminal activity. They play a cat-and-mouse game with bad actors who use a wide variety of financial instruments in sophisticated ways, exploiting the distributed nature of the financial system to perpetrate their crimes. Criminals might, for example, engage in small transactions across many different institutions, or across different accounts in the same financial institution, to mask their activities.

Protecting privacy while advancing security

The global focus on digital privacy in an increasingly interconnected world is a cornerstone of trust, human rights, and individual empowerment. Privacy is mandated by legislation around the world, such as the Digital Charter Implementation Act 2022 in Canada and in the European Union’s General Data Protection Regulation. And of course, banks and other financial services firms also know that customers will vote with their feet if their data is leaked or mishandled.

At its core, privacy is about protecting personal information. And this poses some challenges in fighting financial crime, because it impairs organizations from knitting together a complete picture of what an individual bad actor or a group of bad actors may be doing. The keys are all there in transaction records, account information, customer relationship databases, and so on. But they remain off limits when they are associated with personally identifiable information.

Fortunately, businesses can now attack the problem using novel technologies such as confidential computing and AI that allow multiple parties to safely gain insights from financial data without violating privacy requirements.

Confidential computing and de-identification: New layers of protection

A host of modern, cloud-based capabilities and methods enables this shift. For one, data can be better protected in the cloud with solutions like Azure confidential computing. This unique service encrypts data while it’s being processed, meaning that data is no longer only protected at rest and in transit, but also in use. While in memory, it simply cannot be accessed by cloud operators, malicious administrators, or even privileged software such as a hypervisor.

The root of trust in Azure confidential computing resides in independent hardware. Not even Microsoft operators can access the encryption keys. This is what enables government customers to independently, cryptographically verify the identity and “known good state” of the cloud operating environment they are relying on.

Concurrently, regulators are beginning to recognize the impact of new techniques for de-identification, which obfuscates or removes personally identifiable information from data sets. Data masking, data perturbation, and differential privacy are some of the powerful tools and methods of de-identification that are proving their effectiveness by making data available to AI to deliver important insights without putting privacy at risk.

While securing the benefits of strong privacy protections, financial services organizations are now able to work across enterprise data sets—to reason over data from not just one location, but across different locations and potentially even different institutions. This dramatically changes how a firm handles data. Swift is just one recent example of a financial services firm that has benefited from these innovations in building an anomaly detection model for transactional data without copying or moving data from secure locations. And, significantly, it means that AI and related tools and technologies will now be able to explore, analyze, and spot trends and insights that not only help their businesses, but can have positive societal impact as well.

How AI helps financial services organizations

With AI, financial services firms have new capabilities for risk assessment and scoring, which can help prioritize investigations and resources. They can also benefit from pattern recognition, which can detect anomalies and suspicious activities across large sets of financial transactions, customer data, and other sources. This has significant implications for fraud management, which financial services organizations rely on to mitigate their risks. If a firm can show new levels of due diligence, underwriting costs can potentially be reduced.

Additionally, generative AI can be used to analyze a wide array of unstructured data from a variety of internal repositories to spot indicators of potentially suspicious activities. Natural language processing will assist in the delivery of regulatory documents, legal texts, and compliance reports. And financial institutions may realize broad organizational benefits through integration into productivity applications. At Microsoft, we’re all about democratizing AI and making these tools approachable and available not simply to the data analysts and mathematicians, but to people across the business. This is reflected in the broad innovations announced recently at Microsoft Build 2023, in which we have integrated AI into Azure, Microsoft 365, our development tools, and much more. These AI-powered products help surface more useful information for better decision-making and greater efficiencies across the organization.

The art of the possible

In our work with customers, we see a wave of interest in exploring the potential of these powerful new tools to fight fraud, money laundering, and other forms of financial crime. In Canada, privacy enhancing capabilities have long been bolstered by affirmation from the Information and Privacy Commissioner of Ontario that de-identification is a legitimate and valuable way to protect information, and enterprises have been provided with guidance on how to proceed. It’s powerful confirmation that organizations can leverage new approaches to address privacy considerations as they explore new opportunities. Once we light up the art of the possible, the dialogue quickly shifts and we can work collaboratively to solve these tough challenges.

Fight financial crime with the Microsoft Cloud

Collaboration is the key to industry-wide progress in the fight against all kinds of financial crime and fraud. Working well together is a core Microsoft value, and that means much more than ensuring that our products and tool sets are integrated. It means that we recognize that these challenges are bigger than us or any one company, organization, or entity. So, we promote and support the roles that every player in the ecosystem performs, from industry partners to government officials, regulators, law enforcement agencies, and of course customers.

For financial services organizations who want to explore these new possibilities, an exploratory engagement or proof-of-concept is a good way to examine how the technology and process puzzle pieces fit together. We’re constantly amazed at the inventive and impactful ways that customers are employing these tools to do better for their organizations and the world at large.

Read further in a recent post about how the Microsoft Cloud helps banks manage risk and discover real-world customer examples and other resources that show how Microsoft and our global partners can help banks deepen risk insights, facilitate regulatory compliance, and combat financial crime.

124 Alarming Money Laundering Statistics [New Data 2022 & Infographic], BusinessDIT.

2Global spend on financial crime compliance at financial institutions reaches $213.9 billion, Finextra.

3True Cost of Finacial Crime Compliance Study for the United States an Canada, LexisNexis Risk Solutions.