Artificial Intelligence (AI) is the latest technology to permeate the fast-moving financial sector, finding applications in everything from trading and fraud detection to credit risk evaluation and marketing. The widespread deployment of AI within financial services organisations stands to unlock an additional $1.2 trillion in value by 2035 according to Accenture's report "How AI boosts industry profits and innovation". One business function within the financial sector however has thus far been slow in benefiting from AI’s capability to streamline processes and future-proof organisations. The question of when compliance teams will begin to adopt AI’s success stories as their own has become more urgent as financial services realise compliance directly affects every line of their business.

The AI laggard

While some big players such as Barclays have addressed the opportunity, most compliance functions are wedded to manual processes increasingly unfit for purpose. This against a backdrop of failure: globally, the financial services sector has paid out $342 billion in fines since the financial crisis – a figure expected to rise to $400 billion by 2020. This analysis, carried out by Quinlan and Associates, suggests that thanks to writedowns, trading losses, fines and higher compliance costs, the top 50 global banks have seen $850 billion in profits wiped from their organisations since 2008.

A perfect fit for Compliance

In an environment where regulation is increasing in both volume and complexity and the cost of compliance is shooting upwards, AI offers a near-perfect response. Natural Language Processing (NLP), a subset of AI concerned with the interactions between computers and human languages, allows algorithms to analyse documents written in natural language (rather than translated into rule-based code). The technology allows huge volumes of documents to be ‘read’ by machines and augment or even automate certain processes. The initial applications include AML/KYC client onboarding where identity documents can be checked in seconds and an account opened in minutes. But recently a new wave of AI technologies is targeting other, more complex use cases that go deeper into regulatory documentation and internal policies, procedures and controls in order to help identify compliance gaps.So far, so ‘blue sky’; you might be thinking. In fact, the technology is already widely used. One of the most famous recent applications comes from the OpenAI initiative, which created an AI application that not only reads natural language, but can also write it. Another example can be found within Google, which is developing a natural search engine that can ‘read’ the web and then answer complex questions on any topic. In the financial services sector, few companies have yet grasped the full implications of NLP-based applications for their Compliance functions, especially where new algorithms used by tech giants elsewhere are being applied to compliance. Those that have are realising unprecedented benefits.

AI Compliance in action

A number of banks use the ClauseMatch AI compliance platform, which combines machine learning and NLP capabilities. Banks use our technology to ensure that their regulatory obligations, policies and risk controls they maintain across the businesses are mapped throughout to mitigate the risk of non-compliance. The AI analyses the content of documents to find and suggest relevant regulatory text, policies and controls so that nothing get missed or amended in a contradictory manner. The system also automates policy lifecycle and maintenance, ensuring that policies and controls remain aligned as and when regulatory changes occur. The approach has removed the time and effort associated with compliance for banks and ensures that all its governance documents are not only compliant but fully ready for audits.

Future-proof Compliance

Given the increasing scope of financial regulation it’s already nearly impossible for firms’ compliance teams to map the thousands of policies they have in place against their risk controls and regulatory obligations. AI and automation are the best tools to create a compliance function fit for the future. The adoption of AI will help remove the threat of increasingly punitive financial penalties for non-compliance while helping banks enable some $217 billion of cost efficiencies in their back-office processes.