Artificial Intelligence
If compliance kills, will AI come to the rescue?
- The cost of banks compliance is estimated to be between 11-20 percent of operating costs, a 14 point increase from 2014.
- Streamlining the compliance with AI increases the productivity of compliance officers by about 65 percent.

With every financial crisis, a backlash of regulation ensues. This is especially true for the last financial crisis and the Dodd-Frank Act which followed, considered by many as the most comprehensive financial reform since the Great Depression.
Over 400 new rules were created as part of DFA, including the creation of a new regulator, the Consumer Financial Protection Bureau.
With each wave of regulation, banks' burden of compliance increases. According to KPMG’s 2016 Regional and Community Banking Industry Outlook Survey, 47 percent of banks estimate the cost of compliance between 11-20 percent of operating costs, a 14 point increase from 2014.
Through small banks have received accommodations in recent major rules, they are consistently exiting the game, mostly through M&A, according to a Congressional Research Service paper on regulation effect on small banks.
The baking game is getting harder for smaller fish.
Dale Wilson, CEO of First State Bank in San Diego, Texas, explained this vividly in his testimony to the Financial Services Committee on Dodd-Frank. "During the last decade, the regulatory burden for community banks has multiplied tenfold," he said. "Since the passage of Dodd-Frank there are 80 fewer Texas banks. These banks didn't fail…. [they] could not maintain profitability with regulatory cost increasing between 50-200 percent."
In some aspects, regulation costs hit larger banks harder, but they have the resources to sustain it.
Major ongoing regulatory concerns for banks include Anti-Money Laundering investigations, or Know Your Customer rules, which require banks to gather and infer more information about customers, to determine the risk in onboarding new clients. Banks are also subject to regular compliance reporting and periodic stress tests that check their ability to sustain a crisis.
To comply with onboarding regulation, some banks employ a compliance outsourcing company in a low cost location, where an army of people compare documents in a mostly manual process. In recent years, more financial firms are using artificial intelligence, machine learning and advanced rules to automate large portions of this process.
By automating the process, banks can achieve an order of magnitude change in spending, said Michael Henry, general manager for KPMG’s global automated platform business for compliance and regulatory reporting. Streamlining the investigation process with software that uses some form of artificial intelligence increases the productivity of compliance officers by about 65 percent, according to Geri-Lynn Clark of NextAngles. Such software can process a large number of cases, elevating harder ones to humans, thus achieving significant savings in man hours. One client Henry worked with was able to release 300 people after deploying such a platform.
New AI enables software to be more agile and flexible to adapt to the changing regulatory environment, rather than being a specific solution to a specific regulation. KPMG's Henry envisions the future of compliance activities as a “cold, dark room", containing just one big computer that can produce any report on demand.
Though it might be too late for some smaller banks that were forced to shut down by the burden of compliance costs, big banks with shrinking margins will certainly benefit from new technology.
Photo credit: Steven-L-Johnson via Visual Hunt / CC BY

