The REGTECH Book

Enabling RegTech Up Front: Unambiguous Machine-readable Regulation

By Marcel Fröhlich¹ and Donald Chapin²

¹Director Services, Eccenca GmbH

²Principal Consultant, Business Semantics Ltd

RegTech functionality ranges from data collection technologies to complex decision automation and behavioural analysis. Some tools have an obvious task, e.g. ‘record customer dialogues’, whereas others try to apply various artificial intelligence (AI) approaches to extract definitions and rules out of regulatory documents to automate complex checks.

Our focus is on the latter type of RegTech solutions, which are especially relevant for large regulated organizations like banks operating in multiple jurisdictions. This chapter does not describe a new method of information extraction but rather turns the challenge upside down and tries to answer the question: ‘What can the regulators do to enable RegTech?’ Regulators are the owners of regulatory instruments, typically rules and definitions provided in a hierarchically structured textual form (e.g. the FCA Handbook).¹

Regulations are often a mix of principles-based (outcome-oriented) and rules-based (ticking checkboxes) approaches to define compliance. In general, clear rules are more accessible for automation, but in both cases the key hurdle for automation is clarity of meaning with ambiguity removed. Without unambiguous, easily understood defined meanings for the terms used in regulatory rules, automation is useless because meaningful logical ...