Many regulations impact the way banks and financial services firms make commitments or conduct transactions with partners and customers. This means that they must find relevant contracts, review the affected language, and then make business decisions to revise or novate the contract, renegotiate commercial terms, or terminate to avoid non-compliance.

The Institute of International Finance concludes that the cost of addressing regulatory requirements can reach into the billions of dollars for many institutions. Critically, these costs are not resolved as part of a single event or effort.  Instead, an institution may have to evaluate the same data multiple times to satisfy multiple requests by regulators looking to improve visibility into risk, benchmark against peer institutions, evaluate compliance with ever changing standards, and identify other critical gaps in the financial ecosystem.

Artificial intelligence provides banks and financial services firms with an important tool with which they can address regulatory compliance in a more agile and efficient way. This is easing compliance processes and making financial service providers and banks more nimble and less exposed to the extraordinary costs and delays normally associated with manual contract review, analysis and remediation.

Artificial intelligence is the emerging weapon of choice

To achieve the delicate balance between financial stability and operational efficiency, advances in artificial intelligence now allow banks and other financial institutions to understand both existing and new regulations, and implement appropriate changes by stated deadlines to avoid penalties, fines, or worse. This means these organizations are better prepared to comply with new rules and deal with audits in a way that reduces both cost, effort and time to value for the organization, thus freeing up precious resources for other activities.

Artificial intelligence can extract the specific terms and provisions needed for regulatory compliance from across all contracts. An example of this is an advanced machine-learning framework which can be taught by users to look for specific provisions and clauses without the assistance of data scientists, IT staff or specially trained experts. This type of automated compliance review can be done quickly and at a far lower cost than in the past.

Financial institutions are also using AI to incorporate process walkthroughs into their conventional, established compliance-risk assessments. Regular first- and second-line assessments of inherent risk exposures, and how they affect business processes, can be conducted through a series of automated searches for standard and non-standard clauses contained in document text.

A familiar example is the implementation of a systematic process that deploys an artificial intelligence to flag any significant operational terms contained in contracts, such those subject to SR-14, IFRS 16, GDPR and LIBOR, to the second line for examination and action.

Industry players are now applying AI as a first- and second-line tool for objectively measuring risk, as well. This includes quantitatively measuring assessable risks contained in contract language, markers for risks that are by nature more difficult to quantify, and routine contractual inventory of risk-prone outcomes.

Artificial intelligence, notably, is more and more a factor in guiding M&A and other due-diligence scenario-analysis. Importantly, artificial intelligence provides its users with the ability to go well beyond traditional M&A analysis for a limited scope of topics, and instead add value to complex divestiture and post-merger integration projects, change management initiatives and forward-looking assessments of privacy, and other BAU risk hidden in contracts and agreements.

Regulations requiring an understanding of contracts

The regulations that compel financial institutions to have visibility into contracts run the gamut from Dodd-Frank 165 “Stress Test” compliance (ISDAs and CSAs), “Living Wills” reporting for financial services, European Banking Authority “Write Down” Rules (BRRD), and anti-money laundering provisions.

Qualified Financial Contracts (QFCs), as a familiar case in point, generally contain information on positions, counterparties, contractual data such as governing law and guarantees, and detailed information on collateral. Federal mandates require reporting on key information held in QFCs within a 24-hour period so that the FDIC can make rapid contractual decisions on behalf of the bank in the unlikely event of insolvency. Artificial intelligence has proven effective for automating discovery of QFCs, extracting precise provisions, and analyzing this data in order to  design and structure QFCs reports.

Importantly, the value applicable to the QFC use case can be realized across a variety of initiatives across large financial services organizations, from “critical event” Brexit analysis as to whether continuity can be secured by contracting entities located in different use cases, to the regular and ongoing assessment of high-volume agreements such as NDAs and procurement contracts.

Utilizing machine-learning models, it is possible quickly extract key terms from contracts and related documents across these varied use cases, based on not just the actual words but also intent and meaning. It also allows non-technical staff to easily and intuitively train AI models based on guidance and limited intervention from subject-matter experts on large data sets and then use the refined models on vast quantities of contracts.

Lest one think that developing the ability to deploy AI across these use cases is simply an academic exercise or the next wave in labor arbitrage, Accenture Research estimates that these efforts will allow organizations in the financial services sector that embrace this type of AI to improve profitability by an average of 31 percent by 2035.

Clarity into contracts using artificial intelligence

An important benefit of employing an artificial intelligence in these settings is establishment of common data models and sharing of extracted data across business units and departments. This not only creates synergy in contract analysis and protocols, but also facilitates speedy triage of contracts to save time and money while allowing the institution to be thoughtful and proactive in its approach toward contract management.

All of this begins with the identification of relevant documents across an entire contract corpus. This discovery capability of an AI platform is the engine that finds, converts and ingests the contracts into a database. The analytics capability, an engine that uses machine-learning models to parse the contracts and extract the required information, using both inference as well as straight word-matching, is then applied.

Software packages have been designed to conduct data extraction for specific regulatory scenarios. Machine-learning models have been trained and tuned to M&A, LIBOR, QFCs and a host of other regulatory frameworks as a means to improve negotiating position and de-risk contractual relationships, as well as maximize benefits of contract terms.

These applications of artificial intelligence are, consequently, not only useful for first- and second-line regulatory compliance assessments, but can also be leveraged by banks and financial services organizations to improve efficiency across the entire contractual process from legal operations to procurement to sales.

Relieving the regulatory burden with AI

Contract analytics, tapping into the latest advances in artificial intelligence, provide a way to get ahead of the compliance curve with not only structural efficiency and cost savings, but a fundamentally better quality of oversight that stands to materially de-risk operations. In fact, a report by Cisco indicates that regulatory compliance and risk assessment are areas where such technology will have the most dramatic impact on the future of banking and financial services.

Artificial intelligence has proven already to lessen the burden of regulatory compliance on the enterprise. Risks in financial institutions are driven by the same underlying factors, but the stakes are arguably much higher when considering the less-than-pernicious outcomes such as curtailment of core business activities, restrictions on new ones, and massive fines.

Implementation of AI as a tool for finding and extracting relevant data, analyzing it, and taking decisive action to achieve compliance should and is becoming a necessary part of the risk-view arsenal in the banking and financial services sectors.