When it comes to AI adoption in the financial services industry, increased regulatory rigour can both giveth and taketh away.
Artificial intelligence in the financial services industry has been a much-hyped technology that will dramatically change the workflows and processes across many traditionally human-centered aspects of Wall Street across sales & trading, investing, banking, FX, and compliance. Perhaps the pace of change has been a bit overpromised but AI is nevertheless impacting many verticals with an acceleration of new live applications and technologies. The increasing reach and complexity of regulations are making it more critical for companies to automate a significant percentage of the compliance process.
Since the global financial crisis in 2008-2009, regulators have ramped up reporting requirements to increase transparency and adherence to regulations.
This is not necessarily an impediment to doing business, but has required a significant learning process for banks, regulators and tech firms. Though regulation could be a speed bump towards AI adoption in some cases, it can also create opportunities in the short and long term.
The new explosion of data
Financial institutions today are reporting more information about the strength of their balance sheets, liability and off-balance sheet exposures, liquidity measures, collateral and capital levels, and reporting those details more frequently and retaining more granular history than ever before. MiFID II, perhaps the most sweeping European regulatory requirement to affect the financial industry in decades, will drive the creation of a digital goldmine of trading data. This regulation has created the need to record large amounts of data that is well-defined and structured for regulatory review and sharing across counterparties and trading venues. Under MiFID II, which went into effect in January 2018, firms are reporting various pre- and post-trading data, but also venue of execution, venue of publications, transaction ID code and much more.
Much of the data is new and mining it can help firms create better analytics — determining which venue is better, slippage cost, aggregated market snapshots, liquidation cost and other advanced outputs. Data scientists are grappling with the huge amounts of data that’s available — 50-60 billion points of data a day in trading. Those data points can be used to depict the current state of the markets more accurately in real time and be used for prediction models using AI and other advanced statistical techniques.
An evolving symbiotic relationship between regulators and firms
In addition to spurring an explosion in data, new regulations are also creating opportunities for AI to prove its usefulness — by helping firms comply with regulations.
Financial firms are looking for ways for AI and machine learning to help streamline regulatory reporting and compliance. (Machine learning is a segment of AI based on the principle that given enough training data, machines can learn for themselves.) It’s projected that the growing industry, known as “RegTech” could reduce the cost of regulatory compliance.
The increase in the number of regulations and the complexity of multinational companies have made it difficult just to stay on top of new regulations. RegTech has primarily focused on making reporting procedures — such as know your customer rules, tax reporting, or anti-money laundering rules — easier for firms.
One RegTech startup in particular is developing software that lets banks process tax forms to meet compliance rules in real time rather than batch.
RegTech is also helping firms make sense of regulations and see how it might apply to them. RegTech firms are experimenting with platforms that can streamline regulatory research, for example. These solutions can integrate various regulations directly into compliance workflows — allowing companies to streamline organization structures, policies, and interpretations. RegTech solutions can alert management to gaps in control requiring attention — giving firms a chance to identify problems ahead of time rather than being reactive.
One law firm worked with a tech firm to develop an automated toolkit to help firms assess the impact of MiFID II regulation, allowing firms to filter through thousands of pages of regulation to find the relevant areas to a firm’s business type, clients and products. Firms are also developing chatbots to offer expert advice on a particular section of a regulation.
Firms aren’t the only ones eyeing AI and machine learning, regulators are looking as well. The Financial Conduct Authority, an independent U.K. financial regulatory body, has said it is looking into the possible use of AI and machine learning to enforce regulatory compliance. The FCA has also said it’s looking into making its handbook machine-readable — meaning machines would be able to interpret and implement its rules directly.
As AI evolves, regulations and regulators will play an important role both hindering and facilitating its development at alternate turns.
By pushing financial firms to become more transparent, regulations such as MiFID II are the catalysts for an enormous database of valuable information that can allow firms to automate efficiencies into various processes, refine automated trading for smaller types of trades or write better algorithms.
Firms that look opportunistically and get ahead of new regulations could give themselves a competitive leg up.
